Get documents about "RAND versus Hanushek, educational McCarthyism, & more correspondence
Document Sample
correspondence
RAND Responds also supports our hypothesis. Hanushek
E ric Hanushek’s critique of our
RAND study was rather one-sided
(“Deconstructing RAND,” Check the
raises legitimate methodological con-
cerns about the study. In the end, how-
ever, these concerns are empirical issues.
Facts, Spring 2001). The case for the Raising a potential area of concern is
author’s long-held interpretation that not the same as proof that results are
there isn’t “any consistent, positive rela- biased. There are significant unresolved
tionship between increased resources issues in all studies—about the adequacy
and student learning” is far less airtight of controls for family background, the
than he suggests. possibility of differential bias at different
We propose a new hypothesis: Addi- levels of aggregation, and the use of weak
tional resources yield few benefits for measures of achievement and spending.
nondisadvantaged white students but Future research must address these unre-
do raise the achievement of minority solved issues in order to ensure more
and disadvantaged students. consistency in our experimental and non-
Literature reviews by scholars Alan Also supporting my hypothesis is the experimental measurements.
Krueger, Larry Hedges, and Rob Green- evidence from class-size reduction exper- Regarding our study, Hanushek
wald have arrived at different conclu- iments in Tennessee and Wisconsin. raises two issues that are simply wrong.
sions than Hanushek’s concerning the Both showed significant gains in achieve- Our sample size was 271 test scores
effects of resources on student achieve- ment for students in smaller classes, with across 44 states, not just 44 as Hanushek
ment. A deeper investigation than larger gains among minority and disad- implies. Furthermore, estimates from
Hanushek’s also reveals that the “real” vantaged students. Hanushek, however, our equations show that modest
increase in resources over time was much suggests that Tennessee’s STAR exper- increases in resources (of $500–$750
less than “threefold.” The author unfairly iment may have been flawed. per student) can lead to significant score
compares resource increases from 1960 Barbara Nye and her colleagues and gains (one-third of a standard devia-
to 1995 to National Assessment of Edu- Alan Krueger have tested the Tennessee tion) among disadvantaged students.
cational Progress (NAEP) scores from data, searching for evidence of unbal-
1971 to 1996. Comparing apples to anced attrition out of the control and DAVID GRISSMER
apples shows a twofold gain in resources treatment groups, leakage between the RAND Corporation
from 1971 to 1996. Moreover, Richard groups, and nonrandom assignment of Santa Monica, Calif.
Rothstein and Karen Hawley Miles of teachers.They separately found either no
the Economic Policy Institute have flaws or flaws that did not affect the
shown that adjusting the inflation rate
for the labor-intensive nature of educa-
tion reduces the real increase to 60 per-
results in any significant way. Despite
years of analysis, no one has uncovered
evidence that the randomized experi-
T o be useful, debates about the
strengths and weaknesses of
research, especially when it involves
cent. A significant share of this increase, ment was flawed. important policy questions, need to
Rothstein and Miles show, went toward Hanushek also claims that the class- adhere to the facts. Eric Hanushek’s cri-
special education—the benefits of which size results show that reducing class tique of our study of test scores and the
would not appear in NAEP data. sizes only in kindergarten can raise stu- accountability system in Texas does not
In addition, much of this new spend- dent achievement. The experiment pro- meet this standard (“The Sequel,” Check
ing went toward compensatory pro- vided no direct empirical evidence that the Facts, Spring 2001).
grams for minority and disadvantaged small classes in kindergarten followed by Hanushek says that our analysis
students, for whom there have been sub- large classes in succeeding grades would “ignored student background,” but we
stantial score gains over time. produce sustained effects. However, controlled for students’ racial and ethnic
The aggregate NAEP results cited by there is evidence that those students backgrounds by analyzing results for
the author hide achievement gains of who had only one to two years of small each group separately. Racial and ethnic
0.5–0.7 standard deviation for black stu- classes in grades K–3 did not sustain background typically accounts for a
dents and 0.3–0.4 standard deviation their achievement gains through grade majority of the variance in test scores
for Hispanic students. Disadvantaged 8, while those who had three to four owing to family background. Controlling
white students account for most of the years of small classes were able to sustain for this allowed us to study whether the
0.05–0.10 standard-deviation gains of their achievement gains. narrowing of the gaps among racial and
white students generally. Our analysis of state NAEP scores ethnic groups on the Texas Assessment
4 E D U C AT I O N M AT T E R S / S U M M E R 2 0 0 1 www.edmatters.org
correspondence
of Academic Skills (TAAS) was com- Eric A. Hanushek replies: The most mid-1990s. Moreover, the relationship
parable to changes on the highly interesting aspect of David Grissmer’s between race and other family charac-
regarded NAEP exams. letter is that he makes no effort to teristics varies from one school district
Hanushek suggests that our finding defend his own study but refers to the next, further complicating Klein
that Texas students showed dramati- instead to others’ findings on the et al.’s attempt to reduce the complex-
cally more improvement on TAAS than relationship between resources and ity of social life to a single racial
on NAEP should be dismissed because achievement. I take Grissmer’s silence dimension.
TAAS was aligned with Texas’ own cur- as an acceptance of my conclusions The researchers claim to have used
riculum, whereas NAEP is a “generic that the RAND study shows a weak “all of the schools in Texas” and “all of
test of subject matter” derived from relationship between spending the 1992–1998 data from NAEP,” as if
national content standards. This ration- and student achievement and their inferences come from more than
ale implies that the skills needed to read cannot be used to identify good a few averages. Their entire database
and do math in Texas are fundamen- policy choices. consists of gains in average scores in
tally different from those needed in the The argument that new spending math and reading from three specific
rest of the country. We doubt that most was aimed primarily at minorities and tests—TAAS, the Texas NAEP, and
Texans would agree. Indeed, President disadvantaged students is similar to the national NAEP—for three racial
George W. Bush’s education plan calls for other arguments in the RAND study: groups. In other words, they have only
using NAEP to check the validity of it bears some relationship to fact but 27 data points. Such limited data do
gains on every state’s tests. lacks evidence and convincing analysis. not provide a sound scientific basis for
Hanushek says our results “should It is true that the test-score gap an informed debate on testing. Had
be heavily discounted”because they were between black and white students nar- they really used all Texas schools in
based on “small amounts of imperfect rowed during the 1980s, only to stag- the second part of their analysis, they
data.” Our major findings were based nate in the ’90s. However, no evidence would have seen the expected negative
on analyses of TAAS results for all the shows a parallel shifting of resources relationship between TAAS scores
schools in Texas plus all the 1992–1998 to disadvantaged students just in the and students’ level of disadvantage—
data from NAEP’s large and carefully ’80s. More to the point, the RAND a relationship they did not see with
constructed samples for Texas and the report did not even attempt to analyze their nonrepresentative sample of
nation. racial differences in the way resources 20 schools.
Hanushek claims that our research affect achievement, even though the Their testing and peer-review
would not hold up to a “modicum of NAEP data would have permitted points are, of course, assertions, not
scrutiny” and that it used an “impotent” such an analysis. If such differential facts. If it were known that all of the
research design. Two of the preeminent sensitivity to resources is truly impor- states’ mandated curricula are well
scholars in the field served as external tant, RAND’s neglect to analyze racial tested by NAEP, it would save consid-
peer reviewers for our paper. Both differences in the way resources affect erable money in test development and
strongly encouraged its publication and achievement would bias its results would quell some intense political
endorsed its methodology. even further. debate. Moreover, the quality of the
Hanushek suggests that RAND Grissmer’s reintroduction of underlying peer-review process should
bends its standards when “the sponsor highly selective evidence on the effects be judged by the outcome.
pressures are high.”We received no pres- of extra resources and reductions in
sure, and there was no external sponsor class size provides no more support of Correction: The printed version of Fig-
for our study. Hanushek acknowledges his work here than it did in the origi- ure 1 in Eric Hanushek’s “Deconstruct-
RAND’s “undeniable history of pro- nal RAND report. ing RAND” (Check the Facts, Spring
ducing solid research.”Nothing in his cri- Stephen Klein et al. surely know 2001) contains an error that occurred
tique suggests that our study departed better than to claim that separating in the production process and is not to
from this tradition. scores by race adequately adjusts for be attributed to the author. A corrected
students’ background. Such stereotyp- graph, along with its interpretation,
S TEPHEN P. KLEIN ing belies the important heterogeneity appears in the electronic version of the
L AURA S. HAMILTON within and across each population. To article available at www.edmatters.org
DANIEL F. MCCAFFREY illustrate, the chance that the income and, in more extended form, at
BRIAN M. S TECHER of a random black male would exceed www.edmattersmore.org.
RAND Corporation that of a random white male
Santa Monica, Calif. approached 40 percent during the
www.edmatters.org S U M M E R 2 0 0 1 / E D U C AT I O N M AT T E R S 5
correspondence
report, Stancik acted as true scandal was the readiness of public
prosecutor, judge, and officials and editorialists to condemn these
jury.The accused teach- educators without according them a pre-
ers and principals had sumption of innocence and a fair hearing.
no hearing before an
impartial judge to DIANE RAVITCH
respond to the charges. Brookings Institution
They were not allowed Washington, D.C.
to confront their
accusers or review any Gregory J. Cizek replies: Diane Rav-
of the evidence against itch herself acknowledges that it was
them. Thacher con- Rudy Crew, not Edward Stancik, who
cluded that many of acted as judge and jury. Furthermore,
those who were named knowing that it was funded by the
Educational McCarthyism and disciplined were entirely innocent of New York City teachers union, it is
I n his article “Cheating to the Test”
(Features, Spring 2001), Gregory Cizek
refers to allegations about a cheating
any cheating; some were reinstated
before Thacher’s inquiry.
Stancik also relied heavily on inter-
impossible to cast the Thacher report
as unbiased. It hypocritically criticizes
the Stancik report for being one-
scandal in New York City, supposedly views with young children, many of sided, yet did not present even one
the largest such scandal ever reported. whom were asked to recall testing cir- conclusion in support of Stancik’s
Cizek cites December 1999 charges by cumstances two or more years after the findings. One has to believe that an
Edward Stancik, the city’s special com- events. In some cases, Stancik’s staff unbiased report would have had at
missioner of investigation for the pub- never interviewed other adults who were least one positive finding.
lic schools, that 52 teachers and admin- in the classrooms on testing day.Thacher The Thacher report ultimately
istrators had assisted students in found that many children apparently reduces to complaints about Stancik’s
cheating on standardized tests. confused practice tests with the real test. process, not his findings. I share with
Former New York City schools chan- The Stancik investigation combined Ravitch (and presumably with Thacher
cellor Rudy Crew’s response was to dis- some of the excesses of McCarthyism, and Stancik) an abhorrence for
cipline all those named in the report; where the accused were subject to pub- instances of people being railroaded,
some were fired, others assigned to desk falsely convicted, or
jobs. All were publicly humiliated and wrongly impugned.
had permanent stains on their reputa- We may also agree
tions. that, as with any large-
The United Federation of Teachers scale investigation, bet-
subsequently commissioned an inde- ter procedures could
pendent investigation of Stancik’s alle- have been in place,
gations by private investigator Thomas more evidence could
Thacher, who previously served as have been uncovered,
inspector general of the School Con- more time spent
struction Authority. investigating, and
Thacher found that Stancik’s office more carefully crafted
bypassed basic procedural rights of those conclusions proffered.
accused. The special investigator’s team, The true scandal is
Thacher said, pressured children to say that the message heard
their teachers acted wrongly and ignored lic humiliation without a chance to by many parents and their children is
evidence that pointed toward the educa- defend themselves, with the now-dis- that cheating is easily excused, argued
tors’innocence.Accused teachers’requests credited practice of relying on children’s away on procedural grounds.
to have a union representative present memories to destroy the reputations of
for their interrogations were denied by adult caregivers.
Stancik’s investigators. Some of the Cheating was not as widespread in Evidence on Integration
accused were never interviewed at all.
In short, according to the Thacher
New York City’s public schools as the
Stancik report suggested in 1999. The T erry Moe finds that inner-city white
parents who are opposed to diver-
6 E D U C AT I O N M AT T E R S / S U M M E R 2 0 0 1 www.edmatters.org
sity are especially interested in switch-
ing to private schools (“Hidden
Demand,” Research, Spring 2001). This
could lead to the conclusion, says Moe,
that inner-city whites “see private
schools as a way to avoid integration
with minorities.” But Moe is careful
to note that a more “benign” interpre-
tation exists: that inner-city whites who
choose private schools simply don’t value
diversity as much as do whites who see
diversity as a strong reason to stay in the
public schools.
For some parents, race has been and
will continue to be a significant factor in
the decision to go private. However, as
Moe found and recent research sug-
gests, the prevailing effect of school
choice can be to reduce, rather than to
increase, racial segregation.This appears
to be true in voucher programs targeted
primarily at low-income urban families.
Jay Greene, a senior fellow at the
Manhattan Institute, found that 19 per-
cent of the students participating in
Cleveland’s publicly funded voucher
program attended a private school with
a racial makeup similar to the Cleveland
metropolitan area population, com-
pared with only 5 percent of Cleveland-
area public-school students. Fifty per-
cent of voucher students attended
racially isolated schools, compared with
61 percent of public-school students in
the Cleveland metropolitan area.
In Milwaukee, home to the nation’s
oldest and largest voucher program,
racial integration is significantly greater
in participating private schools than it is
in Milwaukee’s public schools. Only 30
percent of students in religious voucher
schools attend racially isolated schools,
compared with 50 percent of Milwaukee
public-school students. This compara-
tive advantage resulted directly from the
1998 expansion of the voucher program
to include religious schools, according to
my research with George Mitchell.
HOWARD FULLER
Marquette University
Milwaukee, Wisc.
www.edmatters.org
Related docs
