Get documents about "Rising Tide research
Document Sample
research
Rising tide
Critics of school choice have grossly underestimated the
public school system’s ability to respond to competition
T H E M O S T S C AT H I N G C R I T I Q U E O F V O U C H E R
programs and charter schools is that they may bleed tra-
ditional public schools of their best students and most
active parents, leaving the children who are left behind
even worse off. Moreover, as the students leave, taking
their per-pupil funding with them, the public schools will
find themselves stripped of the human and monetary
resources necessary to answer the call of competition.
“Skimming,” the term of art for this hypothetical phe-
nomenon, may lower overall achievement, as the down-
ILLUSTRATION BY CRAIG FRAZIER
ward spiral of the public schools swamps any gains made
by the students who take advantage of school choice.
Market enthusiasts have always argued the very opposite:
that competition will improve by CAROLINE MINTER HOXBY
www.educationnext.org W I N T E R 2 0 0 1 / E D U C AT I O N N E X T 69
Vouchers Spur Public School Gains in Milwaukee (Figure 1) education marketplaces for a sus-
tained period. Students in Mil-
Public schools in areas where many students are eligible for vouchers to attend private schools
made greater gains from year to year than schools elsewhere waukee have been using vouchers
to attend private schools since
Annual increase in test scores
the 1990–91 school year, though
between 1999 and 2000
(in national percentile points)
Schools most affected by vouchers
10 Schools somewhat affected by vouchers only in 1998–99 was the cap on
8 Schools not affected by vouchers
6
8.4**## the number of voucher students
7.1**## 7.0**
4 5.3
raised from 1 percent of the dis-
2 3.7 3.0 3.0**#
trict’s enrollment to 15 percent.
0 1.9* The Milwaukee district loses a
-0.1
-2 significant amount of state aid to
Math Science Language the voucher program, enough at
** Statistically significantly different from the trend for unaffected schools at the 0.05 level.
least to notice if not to elicit some
* Statistically significantly different from the trend for unaffected schools between the 0.05 and 0.15 levels. kind of competitive response.
## Statistically significantly different from the time trend for the somewhat affected schools at less than This study examines the trend
the 0.05 level.
# Statistically significantly different from the time trend for the somewhat affected schools at the 0.05 in student achievement in Mil-
to 0.15 level. waukee schools where large
SOURCE: Wisconsin Department of Public Instruction (various 2000); United States Department of Education, School District Data Book shares of the student body are
eligible for vouchers.
the public schools, just as the entry of attend a private school with those who Both Arizona and Michigan have gen-
Federal Express and DHL into the pack- don’t, show substantial achievement erous charter school laws, approving their
age-delivery market forced the U.S. Postal gains as a result of attending a private applications more easily and funding
Service to lower its costs and offer new school. How competition affects the them more fully than other states. They
services, such as Express Mail. Few ana- students who remain in public schools, were both early converts to the charter
lysts expected the Postal Service to be able however, is a relatively unstudied ques- movement, and some of their public
to compete with its new rivals, yet several tion. In the vast majority of cities and schools are now suffering noticeable
decades later it is a worthy opponent. states, charter schools and voucher pro- enrollment and funding losses as a result
Supporters of school choice believe that grams are either too young or too lim- of competition from the charter school
public school administrators and teach- ited for the public schools to have sector.This study examines achievement
ers would respond with equal vigor to the responded in any significant way. trends in districts and municipalities in
prospect of seeing their students and Only in one city, Milwaukee, and in both states where charters have captured
funding walk out the front door. Their two states, Arizona and Michigan, have significant market share.Taken together,
professional pride and livelihood in jeop- the new choice reforms created truly fluid the findings presented here, from Mil-
ardy,they would work harder,adopt more waukee, Arizona, and Michigan, offer a
effective curricula, hire more talented first glimpse at how public schools are
staff, and turn the district office into more The research is in responding to these new forms of school
of a support center than a maker and choice. They suggest that the fears of a
enforcer of rules.They would be spurred on the question of downward spiral aren’t merely overblown.
to innovate in ways that improve stu- They’re simply wrong.
dent achievement and parental satisfac- whether students in
tion. Competition would be the prover-
bial rising tide that lifts all boats. private schools, using Time-Tested Choices
For the most part, the research is in It is important to recognize, before dis-
on the question of whether students in a publicly funded cussing the effects of charters and vouch-
private schools, using a publicly funded ers on public schools, that these new
voucher or paying tuition, perform bet- voucher or paying forms of choice simply add to the varieties
ter than their peers in public schools after of de facto choice already available in
adjusting for all the background charac- tuition, perform many parts of the country. It is not at all
teristics that affect achievement. Studies unusual for parents in a metropolitan
comparing students in private Catholic better than their peers urban area to be able to choose from
schools with students in public school, among many schools and many school
and students who receive a voucher to in public schools. districts and from a variety of low-cost
70 E D U C AT I O N N E X T / W I N T E R 2 0 0 1 www.educationnext.org
research
EFFECTS OF COMPETITION HOXBY
private school options. These kinds of ans in the Midwest, tend to have devel- every school in the nation were to face a
choice have been around for so long and oped a large market for private schooling, high level of competition both from other
are so universally taken for granted that with many options whose tuition is sub- districts and from private schools, the
we tend not to think of them as choice. stantially subsidized by donations of land, productivity of America’s schools, in
But take a family living in the Boston buildings, and money given primarily terms of students’ level of learning at a
area. They can choose among 70 inde- during the first half of the 20th century. given level of spending, would be 28 per-
pendent districts located within a 30- Private schools in areas whose religious cent higher than it is now. And that is
minute drive of downtown and many populations are relatively young or sup- with a relatively diluted form of compe-
more in the metropolitan area. Towns portive of public schools tend to be less tition; traditional forms of choice do not
looking to attract families and raise prop- competitive; they simply don’t have the provide strong competition because
erty values face clear incentives to safe- endowments or contributions to keep money does not follow students in a
guard the quality of their schools. their tuition low enough to make them direct way. Furthermore, traditional
Other metropolitan areas offer far accessible to most parents. With strate- forms of choice are not available to many
less competition among local districts. gies similar to those used to analyze the families, either because they live in an
A family living in Las Vegas, Miami effects of interdistrict choice on public uncompetitive area or because they are
(where one school district, Dade County, schools, I compared public school per- too poor to move to another district or
covers the entire metropolitan area), or formance in areas where public schools pay private school tuition.
Hawaii (where the entire state is one face strong competition from private
school district) will be served by the same schools with public school performance
school district no matter where they in areas where little competition between New Competitors
choose to reside; the district has a virtual public and private schools exists. One advantage of studying traditional
monopoly over public schooling in the My comparison showed that all forms of school choice is the insight
area. Comparing public school perfor-
mance in highly competitive metropoli-
tan areas with performance in less com- If every school in the nation were to face a high
petitive areas is one way of discovering
how competition affects public schools. level of competition both from other districts
My research shows that metropoli-
tan areas with maximum interdistrict and from private schools, the productivity of
choice elicit consistently higher test
scores than do areas with zero inter- America’s schools, in terms of students’ level
district choice. The 8th grade reading
scores of students in highly competitive of learning at a given level of spending, would
areas are 3.8 national percentile points
higher than those of students in areas be 28 percent higher than it is now.
with no competition; their 10th grade
math scores are 3.1 national percentile schools perform better in areas where they give into how competition unfolds
points higher; and their 12th grade read- there is vigorous competition among pub- over the long run—over the many
ing scores are 5.8 national percentile lic and private schools. Areas with many decades traditional forms of choice
points higher. Moreover, highly com- low-cost private school choices score 2.7 have been in place. Only the short-term
petitive districts spend 7.6 percent less national percentile points higher in 8th effects of competition can be seen
than do districts with no competition. grade reading; 2.5 national percentile through the study of vouchers, charter
In other words, interdistrict competition points higher in 8th grade math; 3.4 schools, and their impact on the public
appears to raise performance while low- national percentile points higher in 12th schools against which they compete.
ering costs—the result predicted by grade reading; and 3.7 national percentile Therefore, for this study it was crucial
market enthusiasts. points higher in 12th grade math. to isolate those instances where com-
School districts face competition not In short, both traditional forms of petition was lively and long-standing
only from other school districts but also choice—choice among school districts enough to potentially provoke a com-
from private schools. Metropolitan areas and between public and private schools— petitive response from the public
vary in this regard as well. Areas with influence public schools in a positive man- schools. As noted earlier, only charter
long-standing religious populations, such ner.To place the influence of competition schools in Arizona and Michigan and
as Catholics in the Northeast and Luther- on school performance in perspective, if the voucher program in Milwaukee
www.educationnext.org W I N T E R 2 0 0 1 / E D U C AT I O N N E X T 71
Competition from Charters Improves dents than the Milwaukee
Public Schools in Michigan (Figure 2) schools. The same is true of
the less-treated schools in Mil-
School districts that lost more than 6 percent of their students to charter schools in Michigan responded waukee: they are richer and
to the competitive threat, by improving their scores in math and reading. have fewer minorities than the
10 most-treated schools. Research
post-reform (in scale points); range
has shown that these richer
Annual increase in test scores
Schools not affected by charter competition
8 schools outside Milwaukee and
running from 0 to 10
Schools affected by charter competition the less-treated schools in Mil-
6 6. 4* waukee are ordinarily likely to
improve more rapidly than the
4
most-treated Milwaukee
4.1*
3.5 3.7* schools. Put another way, the
3.2
2 2.6* rates of improvement in the
2.3
1.6 schools facing the most com-
0
petition from vouchers will
4th grade reading 4th grade math 7th grade reading 7th grade math probably look less impressive
* Difference between these two figures is statistically significant at the 0.05 level. than they actually are. It is sim-
SOURCE: Michigan Department of Education (various 2000).
ply more difficult for perfor-
mance to improve in these low-
met this basic criterion. 1998–99 school year that vouchers gen- income schools than in the less-treated
erated real competition for the public control group schools.
School vouchers in Milwaukee. In Milwau- schools.Therefore, it made sense to com- Given this, the results for Milwau-
kee, students from families with incomes pare public school performance during kee’s most-treated schools are remark-
at or below 175 percent of the poverty line the 1996–97 (before significant competi- able. As shown in Figure 1, 4th-grade
are eligible for vouchers to attend a pri- tion) and 1999–2000 (after significant math scores rose by about 7 percentile
vate school. For every student using a competition) school years. I divided Mil- points per year in the most-treated
voucher to leave the Milwaukee public waukee schools into those that were more schools, 5 percentile points per year in
schools, the school loses state aid equal to “treated” to competition because at least the less-treated schools, and just 4 per-
half the value of the voucher. During the two-thirds of their students were eligible centile points in the control schools.
1999–2000 school year, the year studied for vouchers, and those that were less Social-studies scores in the most-treated
here, the voucher was the lesser of $5,106 treated because less than two-thirds were schools rose by 4.2 percentile points per
or the cost of tuition at the private school eligible. I expected that the more-treated year, while in control schools the scores
the student chose. So the district lost schools would respond more strongly rose by only 1.5 percentile points per
$2,553 of its $8,752 in per-pupil spend- than the less treated, but the latter is not year. The scores of the students in the
ing, or 29 percent, for every student who a true control group. A full 25 percent of most-treated schools, the schools facing
used a voucher. More than 90 percent of the students in the least-treated schools the most potential competition from
the vouchers went to students in grades were still eligible for vouchers. The vouchers, improved by more in every
1–7 during 1999–2000 because the response of these schools to vouchers subject area tested than did the scores of
vouchers were sufficient to cover tuition might be attenuated, but it would still be the students facing less or no competi-
only at private elementary schools; high a response. To find schools to serve as a tion from vouchers. In fact, though read-
schools tend to charge more. Thus the true control group, I used the following ing scores improved only slightly in the
only public schools in Milwaukee that criteria: 1) the schools were not in Mil- most-treated schools (by 0.6 percentile
faced serious competition from the waukee (and so entirely unaffected by points per year), reading scores actually
voucher program were elementary vouchers); 2) the schools were urban; 3) dropped in the less-treated schools (by
schools, so I focused on students’ scores at least 25 percent of the students were eli- –0.4 percentile points per year) and con-
in 4th grade, the only elementary grade gible for free or reduced-price lunch; and trol schools (by –1.4 percentile points
in which all Wisconsin students take a 4) African-Americans composed at least per year), perhaps due to Wisconsin’s
statewide exam. 15 percent of the student body. adoption of a controversial new “whole
It was not until the cap on the voucher Only 12 schools in Wisconsin met all language” reading curriculum.
program was raised from 1 percent of the these criteria. Overall, these schools are Recall that the most-treated schools
district’s enrollment to 15 percent in the still richer and have fewer minority stu- were those with the most to overcome
72 E D U C AT I O N N E X T / W I N T E R 2 0 0 1 www.educationnext.org
research
EFFECTS OF COMPETITION HOXBY
in terms of raising their scores, yet they
bested the less-treated and untreated The mostly poor and minority students in
schools on every measure of improve-
ment. Their mostly poor and minority schools facing the most competition in
students experienced an upward spiral
in achievement as a result of competi- Milwaukee experienced an upward spiral in
tion, and the improvement is even more
impressive than the mere comparison achievement, and the improvement is even
of the numbers suggests. Of course, it
cannot yet be known how long these more impressive than the mere comparison
schools will be able to maintain their
rates of improvement. It is possible that of the numbers suggests.
improvement will slow after a few more
years of competition. have to get their charter approved by imposes few constraints on their growth.
the very district that would be their Consequently, about 5 percent of Ari-
Charter schools in Michigan. Michigan’s competitor. Big-city, small-city, and zona’s public school enrollment attended
charter school program was established small-town schools are all well repre- charter schools during the 1999–2000
in 1994. It was relatively easy to isolate sented among the schools that face char- school year, the highest share of any
the schools targeted by competition in ter competition in Michigan. state in the country. Charter schools in
Wisconsin: no school outside Milwau- As in Milwaukee, achievement Arizona can be state sponsored, in which
kee faced competition from vouchers, improved in Michigan public schools case they get a fee equal to the state’s
and the degree of competition faced by faced with significant competition. As share of revenue (45 percent of the total
Milwaukee schools varied with students’ shown in Figure 2, their scores climbed revenue for a regular public school).
poverty. In Michigan, I had to choose a by 2.4 scale points more per year in 4th They can also be district-sponsored, in
threshold level of charter school enroll- grade reading and 2.5 scale points more which case they get a fee equal to local
ment, above which I would classify the per year in 4th grade math (4th grade is, per-pupil revenue but are less able to
neighboring public schools as being sub- again, the only elementary grade in compete since they must seek renewal of
ject to charter competition. Because which Michigan administers a statewide their charters from the very districts
enrollment in Michigan schools nor- test). These improvements are above with which they compete.
mally fluctuates by about 5 percent a and beyond their achievement trends I followed the same strategy in eval-
year, even when there is no competition before they were subject to charter uating Arizona that I used with Michi-
from alternative schools, I chose a competition. Moreover, they are above gan, with one exception. A single munic-
threshold of 6 percent. That is, I defined and beyond the improvements made ipality in Arizona may contain several
a treated public school as any school in during the same period in public school districts, so rather than associate
a district where charter school enroll- schools that did not face charter com- regular public schools and charter
ment was at least 6 percent of regular petition. Just to give a sense of the mag- schools with a district, I associate them
public school enrollment. nitude of these improvements, one can with a municipality. As in Michigan, I
There are advantages to studying compare Detroit (a district that did used charter school enrollment of 6 per-
Michigan. For one, a Michigan district face competition) with one of its most cent of regular public school enrollment
that loses a student to a charter school affluent suburbs, Grosse Pointe (a dis- as the threshold for classifying a munic-
loses a substantial amount of money— trict that did not face competition). If ipality as facing charter competition.
the state’s minimum level of per-pupil Detroit were to maintain its faster rate Also, as in Michigan and Wisconsin, I
spending, given the characteristics of of improvement, it would close the focused on 4th grade scores because it is
the school’s student population. Dur- achievement gap between its students the elementary grade that has been
ing the 1999–2000 school year, the aver- and Grosse Pointe’s students in just tested statewide for the longest time.
age spending on a charter school student under two decades. The results in Arizona were similar
was $6,600, compared with $7,440 for to those in Michigan and Milwaukee. As
the average public school student. Also, Charter schools in Arizona. Arizona’s char- can be seen in Figure 3, regular public
charter schools in Michigan receive their ter school law, passed in 1994, is widely schools that faced charter school com-
charters from statewide organizations, regarded as the friendliest to charter petition improved both their 4th grade
such as universities. Unlike charter schools. It gives charter schools consid- reading scores and their 4th grade math
schools in many other states, they do not erable financial and legal autonomy and scores by 1.4 national percentile points
www.educationnext.org W I N T E R 2 0 0 1 / E D U C AT I O N N E X T 73
a year. These improvements are above
and beyond their achievement trends The long-term results found in my studies of
before charter competition.They are also
larger than the improvements made over traditional competition among districts and
the same period by public schools that did
not face charter competition. Again, for between public and private schools seem to
perspective, let’s compare a municipality
that did face charter competition, such as confirm that competition is in general good
Phoenix, with its affluent suburbs. If
Phoenix were to maintain its faster rate for the public schools.
of improvement, it would close the
achievement gap between its students the student would be achieving at a
seem to confirm that competition is in
and those in its affluent suburbs in less higher level than he was before vouch-
general good for the public schools.
than ten years. ers induced some students to choose a
In fact, say the critics’ worst fears
new school.
were realized: A school in Milwaukee
If one critique can be leveled against
went from being the best school in the
Conclusion these findings, it is the very opposite of
city to the worst as students began leav-
The findings presented here consistently skimming: that schools subjected to the
ing on vouchers; an extreme creaming
show public schools’ responding favor- most competition might have seen their
effect took place. The top 10 percent of
ably to competition. In Milwaukee, lowest-performing students leave for
Milwaukee schools performed 32 per-
schools facing more competition from charter or private schools. In that case,
centile points better than the worst 10
vouchers improved at rates faster than any achievement gains among the pub-
percent of Milwaukee schools on Wis-
schools facing little or no competition lic schools would be at least partly an
consin’s statewide math exam. Consider
from vouchers. Public schools in Michi- illusion, simply the result of having lost
a student in this hypothetical school: not
gan and Arizona began improving at their worst students to schools of choice.
only the school’s average score but the
faster rates after they lost significant This seems somewhat plausible: the
student’s score on the statewide math
shares of their enrollment to charter worst students might have highly dis-
exam dropped 32 points. Even in this
schools. It is risky to extrapolate from pleased parents who are eager for new
case, so bad as to be barely plausible, if
these short-term results, but the long- options. But the worst students might
the student’s rate of growth in achieve-
term results found in my studies of tra- also have apathetic parents who could
ment were the same as in the Milwau-
ditional competition among districts care less. There is no way of knowing.
kee schools that were most treated by
and between public and private schools Regardless, those who have studied the
voucher competition, after five years
test scores and family back-
Competition Improves Public Schools in Arizona (Figure 3) grounds of students switching
from public to the new choice
Students in districts in Arizona that lost more than 6 percent of enrollment to charter schools began
increasing achievement at rates greater than in schools that weren't affected by charters. schools have found repeatedly
that the students are about
4.0 average, not much different
Schools not affected by charters
from those left behind. Still, if
Annual increase in test scores
Schools affected by charters opponents of school choice
3.0 wish to stipulate that schools
(in percentile points)
3.0*
2.8* of choice actually attract the
2.0 worst students, leaving the
2.0* public schools to teach the so-
1.4*
called cream of the crop, so be
1.0
1.0
it.You’ll get no argument here.
0.8
0.0
0.3 0.3 –Caroline Minter Hoxby is a
4th grade 4th grade 7th grade 7th grade professor of economics at Harvard
reading exam math exam reading exam math exam University and a visiting fellow
* Trend in targeted schools is statistically different from their pre-existing trend at the 0.05 level. at the Hoover Institution,
SOURCE: Arizona Department of Education (1988 through 1995 and various 2000) Stanford University.
74 E D U C AT I O N N E X T / W I N T E R 2 0 0 1 www.educationnext.org
Program on Education Policy
and Governance
HARVARD UNIVERSITY
PEPG
RESEARCH:
An Evaluation of the Children’s Scholarship Fund
Paul E. Peterson and David E. Campbell
An Evaluation of the Florida A-Plus Accountability
and School Choice Program
Jay P. Greene
An Evaluation of the BASIC Fund Scholarship Pro-
gram in the San Francisco Bay Area
Paul E. Peterson, David E. Campbell and
Martin R. West
Available at www.ksg.harvard.edu/pepg/
CONFERENCES:
Testing Testing: MCAS and the Search for School
Accountability
October 11, 2001
Taking Account of Accountability: Assessing Poli-
tics and Policy
June 10–11, 2002
P U B L I C AT I O N S :
The Program on Education Policy and Governance Charters, Vouchers, and Public Education
Paul E. Peterson, Director Paul E. Peterson and David E. Campbell, editors.
The Taubman Center for State and Local Government Brookings, 2001.
John F. Kennedy School of Government
Evidence Matters: Randomized Trials in Education
Harvard University Research
79 J. F. Kennedy Street, Cambridge, MA 02138 Robert Boruch and Frederick Mosteller, editors.
(617) 495-7976 Fax: 496-4428 Brookings, 2001.
pepg_administrator@ksg.harvard.edu
Revolution at the Margins: The Impact of Compe-
tition on Urban School Systems
For more on the activities of Frederick M. Hess
the Program on Education Brookings, 2001.
Policy and Governance (PEPG): The Future of Religious Colleges
Paul J. Dovre, editor.
www.ksg.harvard.edu/pepg/ Erdmans, 2001.
