Martin Carnoy, Rebecca

Jacobsen, Lawrence Mishel,

and Richard Rothstein

Martin Carnoy,

斯坦福大学

(corresponding author)

教授, School of Education

斯坦福大学, CA 94305

电子邮件: carnoy@stanford.edu

WORTH THE PRICE? WEIGHING

THE EVIDENCE ON CHARTER

SCHOOL ACHIEVEMENT

在夏天 2004, the American Federation of
教师 (AFT) published data from the National As-
sessment of Educational Progress (NAEP) showing that
average fourth-grade achievement is higher in regular
public schools than in charter schools, both for students
overall and for low-income students. For black students,
a group that many charter schools are designed to serve,
the analysis showed that average achievement is no bet-
ter in charter schools than in regular public schools.
These conclusions were reported in a front-page article
in the New York Times. Their accuracy has not subse-
quently been challenged.1

Some charter school supporters claimed that the
NAEP data provided only misleading information about
the quality of charter schools because: (1) NAEP only as-
sessed a single year (2003) of fourth-grade scores and so
could not detect whether charter school scores were low
because their students had even lower scores in earlier
grades. 如果是这样, charter school students could have made
more progress even if they still had not caught up to reg-
ular public school students by the fourth grade; (2) 黑色的
and low-income students in charter schools are more
disadvantaged than black and low-income students in

The National Center for Education Statistics (NCES) did subsequently reclassify twelve schools in
its national sample from regular public to charter schools, but this reclassiﬁcation did not affect the
broad conclusions reached by the AFT in its earlier analysis.

C(西德:1) 2006 American Education Finance Association

151

我

D
哦
w
n
哦
A
d
e
d

F
r
哦
米
H

t
t

:
/
/

d
我
r
e
C
t
.

米

我
t
.

e
d
你
e
d
p
A
r
t
我
C
e
–
p
d

我

F
/

1
1
1
5
1
1
6
9
2
9
8
5
e
d
p
2
0
0
6
1
1
1
5
1
p
d

乙
y
G
你
e
s
t

哦
n
0
7
S
e
p
e
米
乙
e
r
2
0
2
3

THE EVIDENCE ON CHARTER SCHOOL ACHIEVEMENT

regular public schools. To make valid comparisons of charter and regular
public school test scores, more demographic information is required than
simply whether students are minority or eligible for free or reduced-price
lunches. Parental education, speciﬁc income levels, and home environment
must also be controlled; 和 (3) charter school performance may be poor
because so many charter schools are new, experiencing “growing pains” or
“shakedown” problems; studies restricted to mature charter schools would
show superior results.

While it is important to analyze these claims, it is curious that the av-
erage underperformance of charter schools came as such a surprise. 这
original intent of charter schools was to allow for experimentation; 前任-
pected outcome of experimentation in any ﬁeld is many failures before suc-
cesses are identiﬁed. While the controversy that erupted in the wake of
the Times article focused on whether students in charter schools, 平均-
年龄, outperform students in regular public schools, it missed a more im-
portant policy discussion regarding the costs of pursuing charter school
policies—weighing whether the successes are worth the failures. The costs
of chartering policies include high student mobility, increased corruption
and mismanagement of some deregulated schools, and more inexperienced
and underqualiﬁed teachers working in schools with fewer hiring restric-
系统蒸发散. Only then will we be able to say whether the underperformance of
some charter schools is a price worth paying for the successes realized by
其他的.

In a New York Times advertisement published eight days later, 并在
several other op-ed articles and web postings, some charter school supporters’
criticism of the use of NAEP data to support claims about charter school under-
performance made a lot of sense. NAEP scores themselves are not a sufﬁcient
basis to conclude that charter school performance is lacking. 尽管如此,
there is extensive corroboration of this ﬁnding.

From NAEP itself, we can reasonably conclude that charter schools do
不是, 一般, serve students who are more disadvantaged than superﬁ-
cially similar students in regular schools. Although NAEP reveals that charter
schools have a higher proportion of black students than regular schools, 黑色的
students in charter schools are apparently less disadvantaged, not more so,
than black students in regular public schools. As Table 1 节目, 尽管 76 每-
cent of black students in regular public schools are low-income, 仅有的 68 每-
cent of black students in charter schools are low-income. In central cities
尤其, black students are more likely to be low-income in regular public
学校 (83 百分) than in charter schools (72 百分). Only in rural schools
are black students in charter schools more likely to be lunch-eligible than black
students in regular public schools.

152

EDUCATION FINANCE AND POLICY

我

D
哦
w
n
哦
A
d
e
d

F
r
哦
米
H

t
t

:
/
/

d
我
r
e
C
t
.

米

我
t
.

e
d
你
e
d
p
A
r
t
我
C
e
–
p
d

我

F
/

1
1
1
5
1
1
6
9
2
9
8
5
e
d
p
2
0
0
6
1
1
1
5
1
p
d

乙
y
G
你
e
s
t

哦
n
0
7
S
e
p
e
米
乙
e
r
2
0
2
3

Mar tin Carnoy, Rebecca Jacobsen, Lawrence Mishel, and Richard Rothstein

桌子 1 Percent of Lunch-Eligible Students, by Race and Location

PERCENT LUNCH-ELIGIBLE

Charter
学校

Regular Public
学校

Difference

全部的

Central city

Urban fringe

Rural

Black Students

Central city

Urban fringe

Rural

–7

–11

–8

–11

–15

笔记: Data are for students who took the NAEP Fourth Grade
Math Assessment and who reported whether they were eligible
for free or reduced-price lunch.

来源: NAEP 2005, supplemented by unpublished data furnished
to the authors by the National Center for Education Statistics.

关于 10 percent of charter school and 4 percent of regular public school
students in the NAEP sample did not provide information about their lunch
eligibility. Under the strongest assumptions about these nonresponders, 这
result holds that urban black students in charter schools are less likely to be
lunch-eligible than their peers in regular public schools. If all nonresponders
in the charter school sample are assumed to be lunch-eligible, and no non-
responders in the regular public school sample are assumed to be, 那么
low-income percentages for black students in charter and in regular public
schools would be 71 和 74, 分别. For black students in central city
charter and regular public schools, the low-income percentages would be 75
和 81, 分别.

Many studies have compared charter and regular public schools at the state
等级, and they usually conﬁrm these NAEP data. We base this conclusion on an
examination of every state-level study we could ﬁnd that had been published or
made available through April 2005.2 一般来说, state studies show that charter
schools have a higher proportion of black students but a lower proportion
of lunch-eligible students than regular public schools. This probably means
that black students in charter schools are less likely to be lunch-eligible than
black students in regular public schools, because black students generally are
more likely to be lunch-eligible than white students. This pattern—more black

2. More detailed descriptions of each of these state studies, including full bibliographic references, 是

provided in our book, The Charter School Dust-Up (师范学院出版社, 2005).

我

D
哦
w
n
哦
A
d
e
d

F
r
哦
米
H

t
t

:
/
/

d
我
r
e
C
t
.

米

我
t
.

e
d
你
e
d
p
A
r
t
我
C
e
–
p
d

我

F
/

1
1
1
5
1
1
6
9
2
9
8
5
e
d
p
2
0
0
6
1
1
1
5
1
p
d

乙
y
G
你
e
s
t

哦
n
0
7
S
e
p
e
米
乙
e
r
2
0
2
3

153

THE EVIDENCE ON CHARTER SCHOOL ACHIEVEMENT

students but fewer lunch-eligible students—characterizes charter schools in
加利福尼亚州, Florida, 伊利诺伊州, 马萨诸塞州, Michigan, North Carolina, 和
威斯康星州. Other state studies cannot conﬁrm this pattern because they did
not collect adequate demographic data, but only one study, that of schools
in the District of Columbia, suggests the opposite—black students are more
likely to be lunch-eligible in charter than in regular schools.

简而言之, the notion that NAEP charter school scores are no better than
those of regular public schools, among all students and particularly for black
学生, because charter school students are more disadvantaged is not sup-
ported by hard evidence.

Many anecdotal accounts, 然而, do suggest that charter school stu-
dents are more disadvantaged than regular public school students who seem
superﬁcially similar. These anecdotes may be accurate, but the data we report
here make sense because such schools are offset by at least as many unno-
ticed examples of charter schools where students are more advantaged than
superﬁcially similar students in regular schools.

And some anecdotes, however well-intentioned, are simply exaggerated.
We examined, 例如, Knowledge Is Power Program (KIPP) 学校, A
highly publicized network of grades 5–8 middle schools serving central city
minority children. Charter school supporters (and KIPP leaders) often claim
that KIPP serves the most disadvantaged students, but careful examination
suggests otherwise. Where fourth-grade test scores are available for a KIPP
school and for the neighborhood schools from which it draws, test scores
of students who transfer to KIPP are consistently higher than neighborhood
averages. Fourth-grade teachers in regular public schools who refer students
to KIPP consistently report that they recommend their most able students, 或者
those with the greatest parental support.

With access to cross-tabulated NAEP data by race, residence, and lunch
eligibility that were not available in August 2004, we can now conﬁrm that
charter school test scores are not higher than those in regular public schools.
As Table 2 节目, black students living in central cities and attending regular
public schools are not only apparently less advantaged than black central city
students in charter schools; they also do better academically, 一般, 在
regular public than in charter schools. This is true in both math and reading
and is statistically signiﬁcant in math.

NAEP data also fail to conﬁrm claims that charter school performance
improves as schools gain experience. In both math and reading, charter schools
that have been providing instruction for four years or more have lower scores
than new charter schools.

These ﬁndings, 也, are conﬁrmed in state studies. Although a few ﬁnd
a charter school advantage in narrowly deﬁned categories (例如,

154

EDUCATION FINANCE AND POLICY

我

D
哦
w
n
哦
A
d
e
d

F
r
哦
米
H

t
t

:
/
/

d
我
r
e
C
t
.

米

我
t
.

e
d
你
e
d
p
A
r
t
我
C
e
–
p
d

我

F
/

1
1
1
5
1
1
6
9
2
9
8
5
e
d
p
2
0
0
6
1
1
1
5
1
p
d

乙
y
G
你
e
s
t

哦
n
0
7
S
e
p
e
米
乙
e
r
2
0
2
3

Mar tin Carnoy, Rebecca Jacobsen, Lawrence Mishel, and Richard Rothstein

桌子 2 Fourth-Grade Test Scores of Black Students in Charter and Other Public Schools by Eligibility for
Free and Reduced-Price Lunch and Location, 2003

MATHEMATICS

READING

Charters

其他
民众

Difference
(Charters
less others)

Charters

其他
民众

Difference
(Charters
less others)

Lunch-eligible

Central city

Urban fringe

Not lunch-eligible

Central city

Urban fringe

Info n.a.

210

208

218

220

217

226

219

212

211

214

227

225

229

219

–2

–3

–6

–8∗

–3

Notes: Data on rural students not available.

∗Statistically signiﬁcant at the 5% 等级.

188

n/a

208

205

198

193

191

196

211

207

214

199

–5

–3

n/a

–3

–9

–1

来源: NAEP 2005, and unpublished data provided to authors by the National Center for Education
统计数据.

California studies ﬁnd that charter schools that converted from regular school
status perform relatively well, but start-up charter schools do not), most ﬁnd
that charter school students perform less well than regular public school stu-
凹痕. States where such is apparently the case include Arizona, 加利福尼亚州, 这
District of Columbia, 伊利诺伊州 (with the exception of Chicago), Michigan, 北
Carolina, and Texas (except for schools chartered by local districts). 其他
状态, charter school students seem to do better in some grades but worse in
其他的.

Seven studies (in Arizona, 加利福尼亚州, Florida, Michigan, North Carolina,
and two separate studies in Texas) compare like students and estimate score
gains for one or more cohorts as students progress from year to year. Three
of these studies (the Florida and North Carolina studies published in this
问题, and one study in Texas) make strong corrections for selection bias. 这样的
studies meet or exceed the quality standard set by charter school supporters
who signed the New York Times advertisement. Six of the seven studies (Arizona
was the exception) estimate that gains were either the same in charter schools
as they were in regular public schools or lower.

总共, the results of state-level studies, including the most methodolog-
ically sophisticated ones, accord with the “no charter school achievement ad-
vantage” inferences drawn from the NAEP data, even though NAEP data were
only for a single year and demographic controls were minimal. The state-
level studies seem strongly to suggest that generally charter schools do not

我

D
哦
w
n
哦
A
d
e
d

F
r
哦
米
H

t
t

:
/
/

d
我
r
e
C
t
.

米

我
t
.

e
d
你
e
d
p
A
r
t
我
C
e
–
p
d

我

F
/

1
1
1
5
1
1
6
9
2
9
8
5
e
d
p
2
0
0
6
1
1
1
5
1
p
d

乙
y
G
你
e
s
t

哦
n
0
7
S
e
p
e
米
乙
e
r
2
0
2
3

155

THE EVIDENCE ON CHARTER SCHOOL ACHIEVEMENT

outperform regular public schools even when the charter schools have had
time to mature and shake out early problems. In a North Carolina study, 什么时候
the strictest controls were used to correct for selection bias by comparing the
same student’s performance when that student transferred back and forth be-
tween charter and regular public schools, the effect on students of being in
charter schools tended to be negative.

Start-up pains might be a valid reason to hold charter schools to a lower
标准, but only if student transiency results from charter schools being new,
not from a tendency of such schools to churn students more than regular public
schools do. If students are new to charter schools (and not yet sufﬁciently at
ease to express their academic potential) not because the schools themselves
are new but because even established charter schools tend to be revolving
doors for students, then student transiency cannot be an adequate excuse for
low achievement, and charter schools should not be held responsible for it.

The authors of several state-level analyses point out that when choice is
made available, school switching is quite common. The most extreme case
is Arizona, where half of all students changed schools during the two years
under study. In Texas and North Carolina, a high percentage of students also
move in and out of charter schools. Students typically do more poorly when
they ﬁrst enroll in a new school. Students who switch frequently may do worse
in “better” charter schools than they do in “worse” regular public schools.

Some might expect charter schools to reduce student mobility, 因为
students can remain in these schools even after families move to different
街区. 相比之下, children who attend regular zoned schools do
not usually have the option to remain in the same school. But this basis for
stability in charter schools can be offset if choice becomes a habit for families
who once exercise it. The data suggest that school shopping by charter school
parents may be a more powerful inﬂuence than the opportunity to remain
in a school of choice when residence changes, so on balance, charter schools
increase student mobility.

Because changing schools has a negative impact on academic achievement,
not only on individual students who change but also on other students in the
学校, a downside of choice may be that the increase in student mobility
causes positive impacts from more school options to be offset by negative
impacts of increased school movement.

One claim made by critics of the AFT’s analysis of the NAEP data was
that although there are no standardized measures of gain scores by which a
charter school’s effectiveness could accurately be detected, this is not a serious
problem because charter schools, unlike regular schools, are shut down if
their student performance is inadequate. But evidence indicates that charter
schools are rarely closed for poor academic performance. When they close, 它

156

EDUCATION FINANCE AND POLICY

我

D
哦
w
n
哦
A
d
e
d

F
r
哦
米
H

t
t

:
/
/

d
我
r
e
C
t
.

米

我
t
.

e
d
你
e
d
p
A
r
t
我
C
e
–
p
d

我

F
/

1
1
1
5
1
1
6
9
2
9
8
5
e
d
p
2
0
0
6
1
1
1
5
1
p
d

乙
y
G
你
e
s
t

哦
n
0
7
S
e
p
e
米
乙
e
r
2
0
2
3

Mar tin Carnoy, Rebecca Jacobsen, Lawrence Mishel, and Richard Rothstein

is almost always because of ﬁnancial mismanagement (going broke or having
funds stolen by unscrupulous charter operators), not because of a failure to
meet academic goals.

The Center for Education Reform (CER), a prominent charter school pro-
moter, published a list of all charter schools that had been closed by October
2002, and the reasons for closure. The center identiﬁed fourteen closures
for academic reasons, representing less than half of 1 percent of all char-
ter schools. In view of the extensive data showing that charter schools often
post lower scores than demographically comparable regular schools, it is im-
plausible that these fourteen schools represent a signiﬁcant proportion of
the low-performing charter schools that should be closed if there were true
accountability for academic performance.

A cursory examination indicates that even the identiﬁcation of fourteen
closures is probably an exaggeration. Arizona has a higher proportion of stu-
dents in charter schools than any other state; CER found two Arizona charter
schools that were closed for academic shortcomings. 一, 然而, voluntarily
returned its charter after only a year of operation, too short a time for charter-
ing authorities to evaluate test scores or other academic data. In the other, 这
school’s chief executive abandoned the school, and enrollment fell below ten
学生. Although the school may also have had academic shortcomings, 它
was its virtual collapse that caused revocation of its charter.

The lack of a charter school achievement advantage suggests that if there are
excellent charter schools that provide better educations than the regular schools
from which students came, there are also many ineffective charter schools
where education is worse. In average data, the better and worse may offset
彼此. NAEP strongly suggests, but does not prove, that most ineffective
charter schools are not being closed, even after academic shortcomings become
obvious.

Just as institutional inertia protects low-performing regular public schools
from reform, it also protects low-performing charter schools from reform or
closure. Charter schools often enjoy the support of organized political, parental,
and community forces; it is much easier for state and district ofﬁcials to ignore
poor performance than to intervene and provoke unwelcome controversy. 这
insistence of some charter school supporters that NAEP data did not reveal real
problems avoids addressing the failure to hold charter schools accountable, 在
view of the inevitable barriers to such accountability that exist in a democratic
political environment.

Many charter school supporters have not resolved a theoretical ambivalence
about whether market forces (parent choices) are sufﬁcient regulatory mech-
万物有灵论, or whether states should ensure that charter schools really do operate
in ways that are likely to, and in fact do, raise student achievement. 和这个

我

D
哦
w
n
哦
A
d
e
d

F
r
哦
米
H

t
t

:
/
/

d
我
r
e
C
t
.

米

我
t
.

e
d
你
e
d
p
A
r
t
我
C
e
–
p
d

我

F
/

1
1
1
5
1
1
6
9
2
9
8
5
e
d
p
2
0
0
6
1
1
1
5
1
p
d

乙
y
G
你
e
s
t

哦
n
0
7
S
e
p
e
米
乙
e
r
2
0
2
3

157

THE EVIDENCE ON CHARTER SCHOOL ACHIEVEMENT

冲突, 反过来, reﬂects (and in some ways distorts) a deeper ambivalence that
runs through the charter school movement. Are charter schools intended to
give parents the ability to choose their own goals, whatever those goals might
是, or should the public insist that public money be spent and charters granted
only to pursue the public’s goal of higher student achievement? If charter
schools’ purpose is the former, then it really shouldn’t matter whether NAEP
scores, or any other measure, indicate that average charter school performance
低; if parents are comfortable in schools that produce low achievement,
they should be permitted to enroll their children in them.

Charter school proponents with this view should not only be unworried
about low NAEP scores but should also be hesitant about the No Child Left
Behind law and its national requirements for minimal academic sufﬁciency. 如果
parental choice were a sufﬁcient guarantee of school quality, then any regular
public school in a district where parents can choose to remain in a zoned
学校 (or switch to a magnet or charter) should be exempt from further
accountability.

The negligible achievement differences between disadvantaged students
in charter and regular public schools also raises questions about early charter
school theory, which assumed that regular public school student performance
was inadequate, especially for disadvantaged students, because union con-
tracts and school district bureaucratic procedures prevent dedicated, creative,
and innovative school leaders from developing new and more effective ways
of running schools. Early charter school advocates argued that, freed from
constraints of bureaucratic rules and union contracts, schools would improve
instruction.

如果, 然而, charter schools are not raising disadvantaged children’s
achievement, the cause of low student performance may not be bureaucratic
规则. When a treatment doesn’t work, it is prudent to examine not only
whether the treatment should be improved, but also whether the diagnosis
might be ﬂawed.

The ﬂaw stems from a failure of charter school proponents to distinguish
between exceptional or anecdotal experiences and the typical experiences
of schools. It is doubtlessly true that creative and effective school leaders,
freed from bureaucratic regulations and union contracts, can design excellent
schools that do a better job of educating disadvantaged children than do typical
regular public schools.

But bureaucratic regulations and union rules do not exist for the purpose of
suppressing creative practices. They mostly aim at preventing corrupt, incom-
petent, and ineffective practices. Freed from regulations, the best educators
can design excellent charter schools. But freed from the same rules, the worst
educators can also design terrible schools.

158

EDUCATION FINANCE AND POLICY

我

D
哦
w
n
哦
A
d
e
d

F
r
哦
米
H

t
t

:
/
/

d
我
r
e
C
t
.

米

我
t
.

e
d
你
e
d
p
A
r
t
我
C
e
–
p
d

我

F
/

1
1
1
5
1
1
6
9
2
9
8
5
e
d
p
2
0
0
6
1
1
1
5
1
p
d

乙
y
G
你
e
s
t

哦
n
0
7
S
e
p
e
米
乙
e
r
2
0
2
3

Mar tin Carnoy, Rebecca Jacobsen, Lawrence Mishel, and Richard Rothstein

Financial management is one area where this has become clear. 许多
school districts are notorious for their bureaucratic mazes, but almost every
rule can be traced to earlier reforms to curb corruption. Because purchasing
department employees were once caught in kickback schemes, districts now
require multiple signatures and reviews of major purchases. School principals
in regular public schools often complain that civil service rules prevent them
from ﬁring janitors whose work is mediocre, but it is also nearly impossible
for them to hire cousins or in-laws for maintenance positions, a practice that
was once routine in public employment.

Charter schools are designed to avoid these rules and thus to enable princi-
pals to hire the most qualiﬁed people and to purchase supplies quickly and at
low cost. 因此, many charter schools can function more efﬁciently. 一些
charter schools spend funds in creative ways that would be prohibited in the
public school bureaucracy: one of the most widely noticed strategies is to hire
younger teachers to work longer hours than regular public school teachers.
Charter schools can pay these teachers more than young teachers in regular
schools but less than typical teachers in regular schools, at the same overall
payroll cost. The strategy requires that many teachers leave as they mature,
but if a charter school can continue to attract young enthusiastic teachers who
are inspired by the challenge, the strategy can be cost-effective.

But many other charter schools, freed from bureaucratic rules that are de-
signed to ensure a minimum level of competence, have developed approaches
that are ineffective. Some charter schools, freed from bureaucratic rules, 是
tainted by corruption and mismanagement, cronyism and nepotism. 尽管
the CER’s survey found that fewer than 1 percent of charter schools had
been closed for academic shortcomings, over seven times that many had been
closed for ﬁnancial or other mismanagement. Freedom from bureaucratic
rules permits some charter schools to be unusually creative and others to
be corrupt or inefﬁcient. Many charter school supporters repeat anecdotal ac-
counts only of the creative schools, but the evidence suggests that these are not
predominant.

Many charter school supporters have seized upon evidence that teachers
are more effective if they attended more selective colleges themselves, 有
higher test scores, or had more college coursework in subjects they were
教学, especially high school science and math. 如果是这样, charter schools could
outperform regular public schools if freed from state teacher certiﬁcation
要求, hiring teachers without formal training in education but with
high test scores and degrees from more selective colleges.

But while some charter schools will hire more qualiﬁed teachers if freed
from certiﬁcation requirements, other charter schools will hire less quali-
ﬁed teachers. They are unlikely to post high performance if their teachers

我

D
哦
w
n
哦
A
d
e
d

F
r
哦
米
H

t
t

:
/
/

d
我
r
e
C
t
.

米

我
t
.

e
d
你
e
d
p
A
r
t
我
C
e
–
p
d

我

F
/

1
1
1
5
1
1
6
9
2
9
8
5
e
d
p
2
0
0
6
1
1
1
5
1
p
d

乙
y
G
你
e
s
t

哦
n
0
7
S
e
p
e
米
乙
e
r
2
0
2
3

159

THE EVIDENCE ON CHARTER SCHOOL ACHIEVEMENT

have neither high test scores and selective baccalaureates nor the pedagogi-
cal training, background in child development, and supervised practice that
traditionally certiﬁed young teachers possess.

Data from the federal government’s 1999 Schools and Stafﬁng Survey
(SASS) 显示, 一般, charter schools have probably not hired more
qualiﬁed teachers. Charter schools were slightly more likely to hire teachers
who had graduated from the most selective colleges (14 percent of charter
school teachers vs. 10 percent of regular public school teachers). But on other
available measures, charter school teachers seem to be less qualiﬁed. 考试用-
普莱, the SASS data show that in mathematics, charter schools were less likely to
hire teachers with extensive mathematics backgrounds. At the secondary level,
where content knowledge is especially important, 仅有的 56 percent of charter
school math teachers had extensive content knowledge in mathematics, com-
pared to 70 percent of regular public school teachers.

For science, charter schools and regular public schools overall hired nearly
the same percentage of teachers with a major or minor in science. 但在
the secondary level, where it most matters, 仅有的 67 percent of charter school
teachers had college majors or minors in science, 相比 78 的百分比
regular public school teachers.

平均而言, teachers typically gain in effectiveness as they gain in experi-
恩斯, up to about ﬁve years, although a few studies ﬁnd effects of experience
that end earlier or later than ﬁve years. The SASS data show that charter
schools have less effective teachers measured in this way. About twice the
proportion of charter school teachers as regular public school teachers had ﬁve
years’ experience or less in 1999. This is a problem not only because of the
inferior instruction that teachers who lack sufﬁcient experience may deliver
but also because the high concentration of inexperienced teachers in charter
schools also deprives these teachers of opportunities for mentoring by more
experienced teachers, one of the most effective ways in which teachers typically
gain skill.

总共, while freedom from certiﬁcation rules undoubtedly permit charter
schools to hire teachers who are more qualiﬁed than typical teachers in regular
public schools, the data do not reveal evidence that charter schools consistently
use their freedom to do so.

The more important question that policy makers should confront is not one
that NAEP data stimulated—whether charter schools, 一般, outperform
regular public schools—but rather whether the underperformance of some
charter schools is a price worth paying for the high performance of others.
How much experimentation should we do on children, knowing that failures
as well as successes may result? This is a much trickier public policy issue,
and it has no easy answer.

160

EDUCATION FINANCE AND POLICY

我

D
哦
w
n
哦
A
d
e
d

F
r
哦
米
H

t
t

:
/
/

d
我
r
e
C
t
.

米

我
t
.

e
d
你
e
d
p
A
r
t
我
C
e
–
p
d

我

F
/

1
1
1
5
1
1
6
9
2
9
8
5
e
d
p
2
0
0
6
1
1
1
5
1
p
d

乙
y
G
你
e
s
t

哦
n
0
7
S
e
p
e
米
乙
e
r
2
0
2
3

Mar tin Carnoy, Rebecca Jacobsen, Lawrence Mishel, and Richard Rothstein

This article is adapted from The Charter School Dust-Up: Examining the Evidence on
Enrollment and Achievement, by Martin Carnoy, Rebecca Jacobsen, Lawrence Mishel,
and Richard Rothstein, published jointly by Teachers College Press and the Economic
Policy Institute (2005). The Economic Policy Institute (EPI) has received unrelated
support from the American Federation of Teachers (AFT); the AFT president is a
member of the EPI board of directors. The research reported in the book and in this
article was not directly funded by the AFT or by any organization with a stake in the
charter school controversy. The ﬁndings reported in the book and in this article reﬂect
the views of the authors alone.

我

D
哦
w
n
哦
A
d
e
d

F
r
哦
米
H

t
t

:
/
/

d
我
r
e
C
t
.

米

我
t
.

e
d
你
e
d
p
A
r
t
我
C
e
–
p
d

我

F
/

1
1
1
5
1
1
6
9
2
9
8
5
e
d
p
2
0
0
6
1
1
1
5
1
p
d

乙
y
G
你
e
s
t

哦
n
0
7
S
e
p
e
米
乙
e
r
2
0
2
3

161
下载pdf