The argument for national curriculum stan-dards sounds simple: set high standards, makeall schools meet them, and watch Americanstudents achieve at high levels. It is straightfor-ward and compelling, and it is driving a seachange in American education policy.

Unfortunately, setting high standards andgetting American students to hit them isextremely difficult. Politically powerful interestgroups must be overcome. Crippling conflictsbetween different religious, ethnic, and ideologi-cal factions must be avoided. And a culture thatis generally averse to an intense focus on acade-mics must be transformed. These challengeshelp to explain why the research on nationalstandards is both very limited and inconclusive.

But what if the research were to clearly showthat having national standards leads to superi-or performance on international tests? Still,there would not be compelling evidence thatnational standards produce optimal outcomes;economic growth, as well as personal fulfill-ment, could very well require an educationfocused on much more than just high testscores.

It appears that the route to successful edu-cation goes in the opposite direction of nation-al standards; it goes toward universal schoolchoice. Only a free market can produce the mixof high standards, accountability, and flexibili-ty that is essential to achieving optimal educa-tional outcomes.

Behind the CurtainAssessing the Case for National Curriculum Standards

by Neal McCluskey

_____________________________________________________________________________________________________

Neal McCluskey is associate director of the Cato Institute’s Center for Educational Freedom and author of the bookFeds in the Classroom: How Big Government Corrupts, Cripples, and Compromises AmericanEducation (Lanham, MD: Rowman and Littlefield, 2007).

Executive Summary

No. 661 February 17, 2010

Introduction

Since at least the publication of A Nation atRisk in 1983, theories have abounded aboutwhat ails American elementary and sec-ondary education.1 One of the major con-tentions has been that the nation’s publicschools simply have not been held to highenough standards, that students are notrequired to clear sufficiently high bars ofknowledge and understanding.

To remedy this situation, advocates of“standards-based reform” have tried to prodstates to set high standards for themselves.More recently, they have championed federalefforts meant to force states to establish stan-dards and tests. The No Child Left BehindAct requires all states to set standards in read-ing, mathematics, and science, test studentmastery of those standards, and get all stu-dents to reading and mathematics “profi-ciency” by 2014.

The effect of these reforms has been spot-ty at best. On their own, states have set stan-dards of widely varying—but generally low—quality. NCLB arguably made matters worse,giving states powerful incentives to set lowstandards and simply label poor perfor-mance “proficient.” That has helped to keepstates and districts out of trouble under thelaw, but has defeated NCLB’s supposedintent: to push standards and performanceto much higher levels.

Confronted with implacable state-levelresistance to high standards, as well as theperverse incentives of NCLB, many stan-dards-based reformers have moved to a newmodel: national standards. In defense of theirproposal, some advocates argue that, in a“flat world” of global competition, it is non-sensical for a single nation to have 50 stan-dards. Two plus two equals four whether achild lives in California, Iowa, or New York.Moreover, they argue that almost every coun-try that outpaces the United States on inter-national assessments has national standards.Finally, they assert that national standardswill introduce desperately needed trans-

parency in American elementary and sec-ondary education. If all states have the samestandards, then the determination of achild’s proficiency will not be based on wherehe lives, but what he actually knows.

These arguments seem compelling, cen-tering on the idea that if you want allAmerican students to do well you simplyneed to set a high national bar and get themover it. Political and educational reality, how-ever, are hardly so simple. Examining thoserealities reveals why states have proven sounfriendly toward high standards on theirown, and why the empirical evidence onnational and sub-national standards is essen-tially inconclusive.

What Is Happening Now

Because of growing frustration withNCLB and escalating concerns about thenation’s economic competitiveness, momen-tum to establish national curricular stan-dards has markedly increased over the lastfew years. Several organizations have pub-lished reports calling for national standards,and many high-profile education leadershave advocated them.2 Recommendationsabout the forms that national-standardsregimes should take have varied markedly—for instance, common standards voluntarilyadopted by states; standards adopted withmonetary incentives from Washington; orstandards with or without accompanyingtests—but the unifying thread has been tohave a single standard applicable nationwide.

The most concrete action has been under-taken by the National Governors Associationand the Council of Chief State SchoolOfficers, which launched their CommonCore State Standards Initiative in April 2009.As of September 2009, 51 states and territo-ries had agreed to support the developmentof “internationally benchmarked” Englishand mathematics standards. If such stan-dards are completed and states choose toadopt them, states will have to agree to makethe standards at least 85 percent of their over-

2

The argumentseems compelling:

if you want allstudents to do

well you simplyneed to set a high

national bar.Political and educational

reality, however,are hardly so

simple.

all English and mathematics standards. TheCCSSI released draft “college and careerreadiness standards” in September 2009, anddraft standards for grades K–12 are sched-uled to be published in early 2010.3

President Obama’s administration hasoffered both strong rhetorical and monetarysupport for adopting CCSSI standards. Wellbefore there were publicly available drafts ofthe standards, U.S. Secretary of EducationArne Duncan suggested that states thatadopt them would put themselves in a betterposition to get part of the $4.35 billion “Raceto the Top Fund”—a chunk of the massive2009 “stimulus” bill—that he controls. Healso announced that up to $350 million ofthe fund would be used to develop assess-ments aligned with the standards.4 And thisis to say nothing of how the standards wouldbe integrated into reauthorization ofNCLB—the core of federal K–12 policy—aprocess that is currently three years overdue.

Why National Standards?

National-standards advocates offer severaltheoretical arguments for their proposals, butalmost all boil down to the same proposition:setting clear, high standards would create asimpler, more coherent, and ultimately moreeffective education system. As a result of hav-ing uniform national standards, the argumentgoes, all students, schools, districts, and stateswould be evaluated on the same metric, andthe ability to hide failings by altering standardsor tests, as is pervasive under NCLB, would dis-appear. A second major argument furnished byproponents is that in the modern world, withits increasingly integrated, globalized econo-my, it makes no sense to continue having mul-tiple standards within one country.

Those arguments notwithstanding, it ispossible that standards advocates could havebeen satisfied with state-level control. Untilrecently, states have, in fact, been the primaryfocus of standards-based reforms, and it isgenerally acknowledged that most authorityover education resides at the state level.5

Similarly, NCLB was intended to push statesto set their own standards and tests, not tohave them designated nationally. But “stan-dards and accountability” advocates havebecome increasingly disappointed by states’standards—and efforts to have all childrenmeet them—leading many to conclude thatstates simply will not force themselves to per-form. As analysts from the pro–nationalstandards Thomas B. Fordham Foundationhave written:

The state standards movement has beenin place for almost fifteen years. Foralmost ten of those years, we . . . havereviewed the quality of state standards.Most were mediocre-to-bad ten yearsago, and most are mediocre-to-badtoday. They are generally vague, politi-cized, and awash in wrongheaded fadsand nostrums. With a few exceptions,states have been incapable (or unwilling)to set clear, coherent standards, anddevelop tests with a rigorous definitionof proficiency. By our lights, you cancount on one hand the number of stateswith clear proficiency standards in read-ing and math and expectations evenapproaching those of the NationalAssessment of Educational Progress.6

States, as the passage suggests, have tendedto set standards low, at least compared to whatgroups like the Fordham Foundation hopefor. It is a situation that has been exacerbatedby NCLB, which leaves it to states to writetheir own standards, prepare and administertheir own tests, and define proficiency forthemselves. The result has been that moststates, to comply with the letter of the lawwhile keeping out of trouble, have set theirstandards very low.

A 2009 analysis that correlated state testscores with performance levels on the NationalAssessment of Educational Progress—a regimeof federal exams used to gauge American stu-dents’ knowledge and skills in several sub-jects—made clear how low and variable statestandards are. On fourth-grade mathematics,

3

Secretary ofEducationDuncan suggestedthat states thatadopt standardswould be in a better positionto get “Race tothe Top” funds.

only one state out of the 48 with available datahad set its proficiency level on par with NAEP’sproficiency level; the rest had set it either at orbelow NAEP’s “basic” level. This was repeatedon eighth-grade mathematics, where only twoof 47 states set their proficiency level at NAEPproficiency. On fourth-grade reading, zerostates out of 48 had set their proficiency levelsequivalent to NAEP’s; a large majority had setit below NAEP’s basic level. The results werejust slightly better in eighth-grade reading;again, no states set their proficiency on parwith NAEP, though more set it equivalent tothe basic level.7

States’ failure to independently set highstandards, and the often huge difference in thedefinition of “proficiency” between states, arepushing standards-based reformers to focuson the national level. By ending what Secretaryof Education Duncan characterizes as thecraziness “of having 50 states designing theirown standards,” states would lose their abilityto “game” accountability by making their stan-dards easier to meet.8

In addition to improving clarity and coher-ence for state-to-state and district-to-districtcomparisons, many in the national-standardscamp argue that we must be able to compareAmerican students internationally. To do that,we must benchmark the achievement of all stu-dents in all states against international stan-dards. As argued in a recent report from theNational Governors Association, the Councilof Chief State School Officers, and Achieve,Inc., “if state leaders want to ensure that theircitizens and their economies remain competi-tive, they must look beyond America’s bordersand benchmark their education systems withthe best in the world.”9 To compete in an inte-grated world economy, they argue, we musthold all students to “world-class” standards.

The First Blush Fades:The Weak Theoretical Case

for National StandardsOn a very basic level, the argument appears

to make sense: set national standards, hold

schools accountable for meeting them, andeducational excellence will ensue. One prob-lem with this logic, however, is that it fails totackle the bedrock question of how to design,implement, and enforce high standards. Afterall, it is getting high standards for all chil-dren—not national standards per se—that isthe ultimate goal of standards-based reform-ers. But as constant state-level failures to cre-ate and meet rigorous standards have madeclear, getting high standards is a very chal-lenging undertaking. Perhaps because of that,most appeals for national standards havebeen couched in superficial assertions about amodern nation needing uniform standards—or algebra being the same in Maine as it is inMississippi—and have stopped there.

That said, a few analysts have attemptedto give more rigorous theoretical argumentsfor centralized standards and testing—though not national standards alone—thanhave its leading proponents.

Economist Robert Costrell has mathemat-ically modeled the potential effects of central-ized education standards. Assuming that (a)the standard is “defined as the required level ofproficiency for a binary credential, such as ahigh-school diploma”; (b) centralized stan-dards-setters are motivated to “maximize theirconception of social welfare”; (c) “utility-maxi-mizing students choose whether to meet thestandard”; and (d) all districts are identical,Costrell asserts that standards will be moredemanding if they are set at a higher, morecentral level than the district. He argues that ina decentralized system, without a way to com-pare districts or schools, districts would gainno advantage in the marketplace by settinghigher standards because high school diplo-mas would not be valued differently byemployers.10 In a subsequent analysis, Costrelladds that a hybrid system of centralized mini-mum standards and a decentralized ability toset higher standards would maximize socialwelfare, leading to an optimal societal combi-nation of leisure and income.11

Of course, modeling is not the same as test-ing with real-world data, so Costrell’s analysisis primarily theoretical. More importantly,

4

Most appealshave been

couched in superficial

assertions about a modern nationneeding uniform

standards andhave stopped

there.

some of his foundational assumptions arefaulty. Costrell is incorrect, for instance, inassuming that employers, as well as colleges,cannot and do not compare districts or jobapplicants’ skills in a decentralized system. Ina recent survey of over 400 employers in theUnited States, almost 45 percent of respon-dents reported that they tested or screened“recent high school and college graduateapplicants to determine proficiency in somespecific basic knowledge/skills (i.e., Math,Reading, Writing, Spoken English, or Other).”12

Similarly, while the college admissions processis often opaque and numerous variables aretaken into account for each applicant, somecolleges do systematically track the perfor-mance of specific high schools’ graduates toassess the standards and performance of thoseschools (and, by extension, their districts).13

What appears to be missing is not an abili-ty of employers and colleges to differentiatebetween districts, but an incentive for districtsto change. Whether they perform well or not,districts typically receive uninterrupted tax-payer funding. The only methods parents andtaxpayers have for punishing unsatisfactorydistricts, as a result, is moving to better dis-tricts or changing authority through school-board elections. But neither of these mecha-nisms can exert much pressure. For the formerto work, dissatisfied constituents must moveto a better district—a costly and inefficient wayto exercise choice. And the latter?

Political realities make it very difficult touse school-board elections to induce change.Consider, for instance, the politicians whowould adopt standards. These actors are asself-interested as anyone else, and as such theirfirst concern is not to maximize studentachievement or social welfare, but to winreelection. And which groups involved in edu-cation are most likely to help the politiciansachieve that goal? While parents and taxpayersare certainly numerous, few are single-issueeducation voters. Most are either concernedwith a variety of political issues or are not verypolitically engaged. In contrast, teachers andadministrators whose livelihoods depend onpublic schooling are highly motivated to focus

on education, and so exert outsized powerover politicians on education issues. Andbecause employees are naturally averse tomanagement raising the standards for theirperformance, it is unsurprising that teachers’unions and administrators’ associations usetheir political influence—which goes farbeyond just voting—to keep standards low.

In looking past the inherent standards-minimizing pressure of democratic control,Costrell is joined by historian Diane Ravitch.Ravitch asserts in National Standards inAmerican Education: A Citizen’s Guide that clearstandards are necessary so that everyoneinvolved in the education process will knowwhat their end goals and measures of successare. She also argues that national standardsmust be established by some government enti-ty lest textbook and test-writing companies setstandards by default. Finally, she contendsthat national standards would help eliminatediscriminatory expectations for studentsbased on race or income level.14 She writes allthis, despite acknowledging the huge strengthof political forces arrayed against high stan-dards. As she wrote of President Clinton’sfailed national standards effort:

I came to realize even as I was writingthat the seemingly straightforward ideaof national standards was falling victimto a tortuous political process. Otheragendas became attached to the basicidea of standards, as legislators and lob-byists from various interest groups sawan opportunity to hang their favoritecauses onto the legislation. The mea-sure placed unnecessary restrictions onhow states could use their tests; dictatedto states the composition of their stan-dards-setting body, ensuring the repre-sentation of every professional groupand mandating affirmative-action crite-ria for selection; insisted on vague andcontroversial opportunity-to-learn stan-dards; and required that any state edu-cational plan encompass all social ser-vices, not just academic standards. . . .Yet I continue to believe that the idea of

5

What is missingis not an abilityto differentiatebetween districts,but an incentivefor districts tochange.

national standards has remarkablevalidity.15

The asymmetry of political influencebetween the public and public-school employ-ees is manifested in the physical presence ofeducation groups in Washington, DC, andstate capitals. Just in and around the Districtof Columbia, the National Education Associa-tion has a headquarters building, and in fiscalyear 2008 employed 896 people on its nation-al staff; the American Federation of Teachershas a headquarters, and in FY 2009 employed409 staffers; and groups such as the AmericanAssociation of School Administrators, NationalAssociation of Elementary School Principals,and the Council of Chief State School Officersalso have headquarters.16 In contrast, the onlygroup ostensibly representing parents, theNational Parent Teacher Association, is head-quartered in Chicago and has only its lobby-ing arm in Washington. And the PTA is hard-ly representative of most parents; it is largelycontrolled by the NEA and has suffered yearsof declining membership.17

These groups, their state affiliates, andtheir members have been very influential infighting efforts to implement rigorous stan-dards and accountability. In FY 2008, the NEAreported spending almost $30 million on“political activities and lobbying,” as well asdelivering over $82 million in “contributions,gifts, and grants” to such groups as theDemocratic Leadership Council and theBallot Initiative Strategy Center.18 The latterdescribes itself as the “nerve center for pro-gressive ballot initiative campaigns across thecountry.”19 In FY 2009, the AFT spent over $24million on politics and lobbying and madeabout $6 million in gifts. Included among theAFT’s gift recipients was the Economic PolicyInstitute, whose “Broader, Bolder” account-ability initiative proposes that states under-take “qualitative evaluation of school qualityand do not rely primarily on standardized testscores to judge the success of schools.”20

Education policy is still, though, primarilymade below the federal level, and state andlocal NEA and AFT affiliates are both numer-

ous and well-funded. In FY 2008, Illinois’NEA affiliate spent more than $1 million onlobbying and other overtly political activities.In roughly that same time, the New YorkState United Teachers—which is affiliatedwith both the NEA and AFT—spent about$4.7 million. The Ohio Education Associa-tion expended almost $3.4 million. TheMichigan Education Association spent about$2.5 million. In the previous year, thePennsylvania State Education Associationexpended nearly $1.9 million on politicalactivities and $372,713 on gifts, including$5,000 to Fairtest, an organization thatopposes the use of “test scores to make criticaleducational decisions about students orschools,” and on its website features endorse-ments from unions around the country.21

Going beyond teachers’ unions, groupsrepresenting numerous other public schoolemployees actively oppose connecting stan-dards to strong accountability based on testperformance. In 2000, Paul Houston, execu-tive director of the American Association ofSchool Administrators, stated that “testingshould be a part of how schools measure stu-dent performance; however, educating stu-dents . . . cannot be measured by one testalone.”22 In November 2006, the executivedirectors of the National Association ofElementary School Principals and theNational Association of Secondary SchoolPrincipals issued an open letter to Congressarguing that school and student progressmust be measured “based on the results ofmultiple assessments and multiple opportu-nities to retake the test.”23 Finally, in 2000, theNational Council of Teachers of English pub-lished a resolution stating that “the use of anysingle test in making important decisions . . . iseducationally unsound and unethical.”24

In addition to powerful resistance fromspecial interests, when accountability mea-sures threaten to keep poorly performing chil-dren back a grade or from graduating highschool, significant pressure to keep standardslow or easily evaded comes from students,families, and sympathetic members of thepublic. In 2007, for instance, when it appeared

6

The asymmetry ofpolitical influencebetween the public

and public-schoolemployees is

manifested in thephysical presence

of educationgroups in

Washington.

that a large percentage of students would notgraduate, Washington State eliminated apending high school graduation requirementthat students pass the mathematics section ofthe Washington Assessment of StudentLearning. That same year, the Maryland StateSchool Board approved an alternate evalua-tion—a project instead of a test—to enablethose who had failed the High SchoolAssessment to graduate.

In the end, all of these political forces haverendered state standards an unappetizinghash. In 2006, analysts at the Thomas B.Fordham Foundation gave standards acrossthe 50 states an average grade of C-minus.Only three states—California, Indiana, andMassachusetts—received As. The problem,they acknowledged, has been that the politi-cal system is stacked against high standardsand tough accountability. Nonetheless,seemingly because they could think of noother option, the Fordham writers endorsednational standards:

We understand that national stan-dards would face the same perils asstate standards. If written by commit-tee, or turned over to K–12 interestgroups, they could turn out to bevague, politically correct, encyclopedic,and/or fuzzy. If linked with real conse-quences for schools, they could bepressured downward. They could evenwind up doing more harm than good.

But if done right, they could finallyput the entire country on the sturdypath of standards-based reform. And ifgreat standards can be written inSacramento or Indianapolis or Boston,perhaps they could be created inWashington, DC.25

Maybe It’s Not the Structure, But the CultureA major factor behind low standards and

poor test results is the great power that thosewho would be held to high standards haveover the American education system. But ageneral American discomfort with centraliza-tion and easily tested knowledge also plays a

part—not something, as will be discussed later,that is necessarily bad.

Many Americans simply do not subscribeto centralized government control of institu-tions, including education, or of test-driven,fact-and-skill-centered learning. Testament tothese proclivities are the popularity of child-centered Montessori and other pedagogicallyprogressive schools, the nation’s long tradi-tion of “local control” in education, and,arguably, the general unpopularity of the NoChild Left Behind Act.26 It is, in fact, quite pos-sible that centralized control of education,desire to perform well in easily tested subjectsand on standardized tests, and focusing onacademic pursuits at all are broadly at oddswith American culture. Centralized control ofeducation, for instance, goes against the grainof what sociologist Seymour Martin Lipsetfamously labeled “American exceptional-ism”—a culture grounded in individual libertyand laissez-faire governance.27 Similarly, thehard-work but anti-academic ethos of thecolonist, frontiersman, and entrepreneur arecentral to the American narrative.

Just as culture in the United States mightbe generally averse to centralization and stan-dardized education, other cultures mightgravitate heavily in the opposite direction.This seems very much the case for East Asiannations, which typically dominate interna-tional tests.

Historian J. M. Roberts notes that in China“an enlarged bureaucracy was to survive manyperiods of disunity . . . and remained to theend one of the most striking and characteris-tic institutions of imperial China.” And howdid one enter the bureaucracy? “The officialswere in principal distinguished from the restof society only by education.” Roberts alsoreports that despite “rich variety in cultureand custom . . . all the East Asian peoples aresimilar in seeming to show great industry andenterprise, and a willingness to accept amarked subordination of the individual tothe group.”28

There is evidence within education thatthere is, indeed, a significant cultural differ-ence between American and other students.

7

Many Americanssimply do notsubscribe to centralized government control of institutions.

American schools and students tend to focusmore on “critical thinking” and other less-concrete and measurable outcomes thanmathematical and scientific skills and knowl-edge. Americans also tend to put much lessemphasis on schooling and academics thanthe people of other industrialized nations,and much greater emphasis on extracurricu-lar activities and part-time employment.29

Specifically looking at elementary and sec-ondary education, Stevenson, Chen, and Leeexamined the attitudes toward education ofchildren and parents in Minneapolis,Minnesota, Sendai, Japan, and Taipei, Taiwan,in 1980, 1984 and 1990. In all three years,despite significant publicity about the highachievement of Japanese and Chinese stu-dents and the low standing of Americans, theresearchers found that Chinese and Japaneseparents were much less satisfied with theirschools than were their American counter-parts. Similarly, they found that Chinese andJapanese parents and students stressed theimportance of hard work to achieve academicsuccess, whereas Americans attributed acade-mic excellence much more to innate ability.30

Finally, there is the difference in academicperformance between Asian students in theUnited States and other U.S. racial/ethnicgroups. Understanding that “Asian” is a des-ignation that includes people as diverse asethnic Japanese and Hmong, Asian studentsoutperform all other U.S. students by sizablemargins, even after adjusting for socioeco-nomic status, on mathematics assessments.31

And since all these groups are in the sameeducation system, this strongly suggests thatculture is a powerful force regardless of theeducational structure.

Paltry DirectEmpirical Evidence

As the Fordham Foundation admits, thereis little theoretical reason to believe thatnational standards will be any better insulatedagainst downward political pressures thanstate standards. Still, national-standards advo-

cates frequently suggest that, theory or no the-ory, national standards are achievable and willwork. After all, as American Federation ofTeachers president Randi Weingarten has writ-ten, “the countries that consistently outper-form the United States on international assess-ments all have national standards.”32

Unfortunately, this “all who beat us havenational standards” factoid is close to theonly empirical support that national-stan-dards advocates typically offer for their cause.As a result, one cannot help but conclude, asshall soon be seen, that the scientific evidenceon national standards is thin and furnishesinsufficient support for the notion thatstrong national standards are either achiev-able—or desirable—in the United States.

Before addressing the empirical literature,it is worth dismissing the factoid. It is truethat most nations that have outperformedthe United States on such tests as the Trendsin International Mathematics and ScienceStudy and the Program for InternationalStudent Assessment have national standards,but so do most nations that have done worse. Toillustrate, on the 2007 eighth-grade TIMSSmathematics assessment, the eight countriesthat outperformed the United States hadnational standards. But, then, so did 33 ofthe 39 nations that scored lower. Moreover,11 of the 12 lowest performers had nationalstandards.33 When looking only at countriesbelonging to the Organization for EconomicCooperation and Development—generally,economically advanced nations—the samenoncorrelation holds: four OECD membersoutperformed the United States, six didworse, and all but the United States andAustralia had national standards. All of thisholds true for the 2007 TIMSS eighth-gradescience assessment, on which all 10 nationsthat outperformed the United States hadnational curricula—but so did 33 of the 37lesser performers and the 9 lowest performers.Among OECD members, five posted betterscores than the United States, five did worse,and only the United States and Australia didnot have national standards.

The most recent TIMSS to include the

8

On the 2007eighth-grade

Trends inInternational

Mathematics andScience Study

(TIMSS) mathematics

assessment, 11 ofthe 12 lowest

performers hadnational standards.

final year of secondary school—essentially,high school seniors—was the 1995 adminis-tration. Very few countries of the 21 partici-pating had adequate student samples toreach firm conclusions, but the results againsuggest that having national standards doesnot correlate strongly with performance. TheUnited States finished poorly—fourth fromlast on the combined mathematics and sci-ence literacy scale—but the three nations itoutperformed all had national standards.Meanwhile, among the top five finishers,three did not have national standards,including the top-finishing Netherlands.34

The 2006 Programme for InternationalStudent Assessment science exam results,which tested 15-year-olds, exhibit similar pat-terns. In 2006, 57 nations participated in theexam, and PISA reported whether 56 hadnational, regional, or fully decentralized sci-ence standards and testing (France had insuf-ficient data). Of the 27 nations with sufficientdata that outperformed the United States, 17had national standards and examinations, 3had external standards and examinations insome regions, and 7 had no centralized stan-dard—very much a mixed bag. The same wastrue for the 28 nations that did worse, thoughthere were proportionately more countrieswithout national standards: 12 had nationalstandards and tests, 3 had regional, and 13had no centralized standards and tests.35

Importantly, few of the nations in the bot-tom half of the PISA standings were devel-oped nations. How did the OECD-onlybreakdown look? Of the 19 OECD nationswith available data that outpaced the UnitedStates, 11 had national standards and tests, 3had regional, and 5 had no centralized stan-dards. Of the 9 that did worse than theUnited States, 4 had national standards andtests, 1 had regional, and 4 had none. Clearly,having national standards is no guarantee ofsuperior performance.

What about the influence of socio-eco-nomic status, which is a very powerful predic-tor of academic success? Of the five nationshighlighted by PISA for having above-averageperformance in science coupled with below-

average impact of socioeconomic back-ground, three had national standards andtests, while two—Australia and Canada—hadregional standards like the United States.36

All of this said, we have testing and stan-dards data for a relatively small number ofnations, and this has made it difficult todraw universally applicable conclusions fromcomparing nations with and without suchstandards. However, some recent empiricalwork has been interpreted as supportingnational standards.

The first major such analysis was conduct-ed by economist John H. Bishop in 1996,though, importantly, he looked at nationalstandards for secondary students coupledwith “curriculum-based external examina-tions” (CBEEs) that students must pass inorder to graduate.37 So far, such testing hasnot been proposed under leading nationalstandards efforts in the United States. Inaddition, Bishop explored the effect of theseexams for secondary students on the standard-ized testing results of 13-year-olds, probingfor a systemic effect of CBEEs.

After controlling for such factors con-tributing to academic achievement as thenumber of books in students’ homes, numberof siblings, and socioeconomic status in 15countries, Bishop found that non-CBEEnations performed worse on the 1991 Inter-national Assessment of Educational Progressthan CBEE nations. Non-CBEE nations hadmathematics scores about two U.S. grade lev-els below CBEE countries, and the resultswere statistically significant. Science resultswere not statistically significant. (Bishop alsoincluded results for geography, which are notdiscussed here because that subject is not afocus of U.S. national standards efforts.)38

There were numerous problems withBishop’s study—indeed, Bishop himself stat-ed that “the power” of his test was “verylow.”39 For one thing, only for the UnitedStates and Portugal, and only in mathemat-ics, did Bishop find lower performances thatwere statistically significant for specific non-CBEE countries, and that significance was atthe borderline 10-percent level.40 Perhaps

9

Among the topfive finishers,three did nothave nationalstandards,including the top-finishingNetherlands.

more importantly, Bishop acknowledgedthat his findings could be spurious, indicat-ing not that CBEEs drive higher achieve-ment, but that nations with CBEEs mighthave cultural or ideological predispositionsthat make them both more likely to haveCBEEs and to do well on achievement tests:

Causation is not proved . . . becauseother explanations for the U.S.,Spanish, and Portuguese lag can nodoubt be proposed. Other sources ofvariation in curriculum-based examsneed to be analyzed. Best of all would bestudies which hold national culture constant.[Italics added.]41

Lending credence to the possibility that cul-ture drives both centralization and perfor-mance, Bishop found that the positive effectof being from an Asian nation was typicallygreater than the effect of having a CBEE.

In 1997 Bishop released a new study thatincluded his IAEP findings. He added ananalysis using 1994–1995 TIMSS mathemat-ics and science data for 13-year-olds in 37nations.42 Regressing participating nations’TIMSS scores against per capita gross domes-tic product, a dummy variable for East Asiannations, and a dummy variable for CBEE,Bishop found that the presence of a CBEE wasassociated with superior performance of oneU.S. grade level in mathematics and 1.2 gradelevels in science. Both findings were statistical-ly significant, though the former was only atthe 10 percent level.

Though more extensive than his 1996work, Bishop’s 1997 findings suffer from thesame problems. Again, Bishop looked not atthe effect of national standards, but stan-dards coupled with high-stakes tests. Andagain, aside from creating a dummy variablefor East Asian nations, Bishop did not controlfor cultural or other unobserved variations.That said, there was very suggestive evidencethat culture might be a highly significantforce behind both the creation of nationalstandards and achievement: in both the IAEPand TIMSS mathematics results, being from

an East Asian nation had a greater positiveimpact than having a CBEE. Indeed, onTIMSS the effect was almost three times morepowerful.

In a 1999 study Bishop again attempted todetermine the effect of national standardsand CBEEs on academic outcomes, this timelooking at the effect of tests controlled atnational and “provincial” levels.43 The studyincluded TIMSS data for two more nationsthan in 1997. Bishop found that the presenceof high school CBEEs improved scores aboutone grade level for 13-year-olds in math and1.3 grade levels in science. Bishop also usedscores on the 1990 International Associationof the Evaluation of Educational Achieve-ment literacy test, which included 25 nations,and the 1991 IAEP tests of 15 nations. Those,too, showed a sizable CBEE advantage,though it was only statistically significant inmath for the IAEP.

In the 1999 study, Bishop attempted tomore directly address the question ofwhether culture might drive both the estab-lishment of CBEEs and higher achievement,and potentially create the spurious impres-sion that CBEEs improve achievement.Unfortunately, he only tried to control forculture in analyzing Canadian provinces,which is not directly applicable to nationalstandards. Moreover, he argued only thatbecause provinces with CBEEs did not havelower rates of disciplinary problems or absen-teeism, they likely did not have greater inher-ent interest in education than non-CBEEprovinces. But provinces in which citizenshave a greater focus on education mightdemand more forcefully that schools fullyreport problems, and those provinces mightdo better academically despite having just asmany kids causing trouble. That is likely onereason Bishop conceded again that studiesadjusting for national culture were needed.44

Following Bishop, economist LudgerWoessmann tackled the effect of standardscoupled with high-stakes assessments. LikeBishop, Woessmann found a strong relation-ship between the presence of CBEEs andsuperior results on TIMSS for 13-year-olds.

10

Nations with curriculum-based

external examinations

(CBEEs) mighthave

predispositionsthat make them

both more likelyto have CBEEs

and to do well on achievement tests.

Using 39 countries that participated in the1995 TIMSS; 38 countries that participatedin the 1999 repeat test; identifying whetherthe nations had CBEEs in mathematicsand/or science; and controlling for 48 “fami-ly, resource, teacher, and institutional controlvariables,” Woessmann found that studentsin CBEE nations scored about 43 percent ofa standard deviation better in math andalmost 36 percent of a standard deviationbetter in science.45

Woessmann’s findings were similar toBishop’s, but so were his study’s shortcom-ings. Most importantly, Woessmann did notcontrol culture adequately to rule it out as amajor unobserved driver of both CBEEs andhigh achievement. Arguing that “concernsabout cultural differences generally arise incross-regional comparisons . . . but should notbe as large within regions,” Woessmann creat-ed dummy variables for regions of the worldto isolate effects within regions.46 However, itis a questionable assumption that the differ-ences among nations within a region are notnearly as great as those between regions.Looking at the variation among nations with-in a region suggests potentially large gulfs.Iran and Israel hardly have the same culture,but they are both in the Middle East. Japanand the Philippines are both in Asia. Spainand Denmark are both in Europe. Finally, forall intents and purposes when looking atTIMSS data, many of the “regions” identifiedby Woessmann are just a few nations. Whilehis Middle East consists of 4 nations, Asia 9,Eastern Europe 13, and Western Europe 20,North America, South America, Oceania, andNorth Africa consist of only 2 nations, andSouth Africa only 1.47

That said, in a working paper very similarto his published analysis (with similar data,controls, and findings), Woessmann listedthe coefficients for each region (he did notpublish them in his final version). Thosecoefficients indicated that there very wellcould be a cultural effect. Like Bishop,Woessmann’s data showed a very strong, pos-itive Asian effect in math, with the Asia dum-my variables providing a boost more than

twice as large as having a CBEE. The EasternEuropean dummy variables also had a sizablepositive effect, coming close to being twice aslarge as the CBEE effect. There were no sta-tistically significant positive regional effectsfound in science, but there were significantdisadvantages in science for the Middle Eastand Southern and Northern Africa.48

In 2004, German researchers HendrikJürges and Kerstin Schneider attempted todetermine the causes of big differences in aca-demic achievement between otherwise similarnations, and included an analysis of CBEEs.Where they differed from Woessmann andBishop was in using only nations belonging tothe OECD. In so doing, they got results verymuch at odds with Bishop’s and Woess-mann’s: CBEEs had negligible effects onTIMSS math achievement.

Why the difference? For one thing, theauthors note, Bishop included greater num-bers of less-developed nations in his analysisthan they did. In addition, Bishop includedsome highly distorting outliers, especially theCBEE-less Philippines, which Jürges andSchneider found to be “a highly influentialobservation.” After adjusting for outliers,Jürges and Schneider concluded that “centralexit examinations do not have a significantimpact on the achievement of middle-schoolstudents.”49 Jürges and Schneider did, howev-er, find one thing that was very similar toBishop’s results: a highly positive effect forAsian nations, suggesting, once again, thatculture could have a powerful influence onachievement.

The most recent study touting nationalstandards comes from the Thomas B.Fordham Foundation.50 In it, Michigan StateUniversity professor William H. Schmidt andhis coauthors argue that several countrieshave instituted national standards—includ-ing Germany, which the paper dubiouslystates has a federalist tradition like theUnited States—and that the United Statescould, too. The paper offers no systematicevidence that as a result of having nationalstandards countries with such standards out-perform the United States, but asserts it

11

After adjustingfor outliers “central exitexaminations[did] not have asignificantimpact” onachievement.

nonetheless. This despite Schmidt havingconceded, when asked at a conference pre-ceding publication of the paper what empiri-cal evidence demonstrates the beneficialachievement effects of national standards,that “it is very difficult to establish that,because virtually everybody in the world hasnational standards.”51

Thin Indirect Empirical Evidence

The scientific literature on national stan-dards is too thin to support any U.S. policymove in that direction. Moreover, whatresearch does exist is handicapped by thepotentially confounding effects of unob-served variables such as culture, and focusesnot just on national standards, but on stan-dards coupled with high stakes for students.It simply does not demonstrate that nationalstandards drive superior performance onstandardized tests. Perhaps, though, theresearch on sub-national standards is moreilluminating.

Bishop supplements his analysis of nation-al-level data with comparisons of academicoutcomes in Canadian provinces, which, likeU.S. states, have autonomy to set (or not set)their own standards. This is importantbecause, with no national standards, Canadahas done very well on international compar-isons, especially PISA. In 2006, Canadian 15-year-olds finished third out of 57 nations inscientific literacy, fourth in reading, and sev-enth in mathematics.52 Of course, this doesnot prove that lacking national standards dri-ves superior performance any more than high-performing countries having national stan-dards proves the opposite. It is also importantto remember that performance data comefrom specific tests, and strictly speakingdemonstrate only the correlation betweenhaving standards and performance on thosetests. And, as shall be discussed further, testresults do not capture all, or perhaps evenmost, desired educational outcomes.53

In his 1996 work, Bishop compared the

IAEP scores in nine Canadian provinces andthe entire United States, adjusting for severalvariables, including the presence of a CBEE.Four provinces—Alberta, British Colombia,Newfoundland, and Quebec—had a CBEE,while the remaining five and the UnitedStates did not. Bishop again found that hav-ing a CBEE led to superior performance, onthe magnitude of between two-thirds andfour-fifths of a U.S. grade-level in mathemat-ics (depending on whether some variableswere counted as influenced or not influencedby CBEEs) and two-thirds of a grade-level inscience.54 Bishop reiterated the findings inhis 1997 and 1999 work.

Despite his findings, Bishop’s analysis ofCanadian provinces and the United Statesfailed again to demonstrate that centralizedstandards drive superior academic perfor-mance. Once again, he assessed the effect notjust of centralized standards, but standardsattached to exams with high stakes for stu-dents. And once again, Bishop did not con-trol for differences in culture, attitudestoward education, or other potentially con-founding variables.

Along the lines of Bishop’s provincialassessment, several researchers have tried todetermine the impact of state-level standardsand tests on academic outcomes within theUnited States. The results, however, have beenas problematic as the evaluations of Canadian-province and national-level standards.

Perhaps the most publicized research onstate standards was a series of studies byArizona State University professors AudreyAmrein and David Berliner. Looking at statestandards and testing coupled with high-stakes for either schools, students, or both,Amrein and Berliner determined that high-stakes accountability in elementary and mid-dle schools was associated with no systematicincrease or decrease in academic achievementas measured by NAEP. At the high school lev-el, implementation was associated withdecreasing ACT, SAT, and Advanced Place-ment scores.55

Amrein and Berliner’s results garnered sig-nificant attention, but suffered from serious

12

The scientific literature on

national standardssimply does not

demonstrate thatnational standards

drive superiorperformance.

methodological flaws. University of Illinoiseducation professor Barak Rosenshine noted,for instance, that Amrein and Berliner com-pared results in states with high-stakes sys-tems against national averages—a mistakegiven that national averages included high-stakes states. Instead, comparing the perfor-mance of high-stakes states against low- orno-stakes states, Rosenshine found that high-stakes accountability had a moderate-to-largepositive effect. He also noted that Amrein andBerliner considered smaller than national-average increases on state NAEP scores to bedecreases. Looking at absolute changes,Rosenshine found that only two high-stakesstates saw scores declines, versus 10 low-stakes states.56

In response to Rosenshine, Amrein andBerliner compared high-stakes to non-high-stakes states and found that high-stakesstates were outperforming low-stakes states.However, the results were only statisticallysignificant on fourth-grade mathematics,not fourth-grade reading or eighth-grademathematics. They also argued that studentswere being exempted from tests at greaterrates in states with high stakes than thosewithout, and that that was a greater explana-tion for rising scores than the presence ofhigh-stakes tests.57

Around the time of Rosenshine’s exchangewith Amrein and Berliner, Hoover Institutionresearchers Margaret Raymond and EricHanushek published a blistering treatment ofAmrein and Berliner’s study. They lodged thesame complaints as Rosenshine, but also tookissue with Amrein and Berliner’s disregardingscore increases in states in which exclusionrates increased, something Amrein andBerliner did on the premise that exclusioncontaminated the results. Reassessing thefindings, Raymond and Hanushek foundthat high-stakes accountability had a signifi-cant, positive effect.58 However, they providedno state-by-state data to support their find-ings, nor did they control for unobserved vari-ables such as culture, ideology, or potential“shocks” such as sobering reports about acad-emic outcomes that might have changed peo-

ple’s attitudes toward education. In 2002, economists Martin Carnoy and

Susanna Loeb released a study on the effect ofexternal accountability on student outcomes.Comparing changes in the percentage of stu-dents scoring “basic” or above on NAEP tothe strength of accountability (roughly, thedegree to which stakes were attached to test-ing outcomes), the authors found that statesthat had implemented stronger accountabili-ty in the early 1990s saw greater fourth- andeighth-grade mathematics gains between1996 and 2000. They also provided a glimpseinto the possibility that variables such as cul-ture or ideology might drive both centraliza-tion of accountability and achievement.Concerning centralization of standards andaccountability, they reported that more gov-ernmentally centralized states, especially inthe South, tended to have higher-stakesaccountability, as did “more populous stateswith correspondingly larger absolute num-bers of disadvantaged minorities, largerschool systems, and larger cities.”59 Concern-ing achievement, they found “no relationshipbetween test-score gains prior to implementa-tion and the strength of accountability,” hint-ing that a cultural or ideological proclivitytoward high achievement did not drive scores.The period “prior to implementation,” how-ever, was just a few years—too short to reliablyidentify a state’s culture or potential effects ofchanging attitudes toward education.60

Perhaps the most important data inCarnoy and Loeb’s study were not central ten-dencies, but graphs of state-by-state resultsfor eighth-grade mathematics. Though“tougher” accountability (measured on a 0 to5 scale) was correlated with better outcomes,the variation from state to state was consider-able, illustrating the pitfalls of assuming thatcentral tendencies are universally applicable.For white students, while the 14 lowest-stakesstates (out of 37 total) tended to have thesmallest gains, three of those states outpacedfive higher-stakes states and tied five more.And the overall outcomes for states in thethree highest accountability levels were heavi-ly influenced by just two very high perform-

13

Outcomes forstates in the highest accountabilitylevels were heavilyinfluenced by justtwo very highperformers.

ers: Louisiana and North Carolina, which saw15 and 14 percentage-point increases, respec-tively. (There were also two states in the high-est accountability levels—California and NewMexico—that saw zero gains.) Were Louisianaand North Carolina eliminated, the change inthe percentage of white students scoring“basic” for states in the top three levels woulddrop from 6.5 points to 5.2. Meanwhile, theaverage increase for states between 1 and 3 onthe accountability scale was 5.4 percent.Results for black and Hispanic students weresimilarly volatile.61

In January 2004, again expanding on thework of Amrein and Berliner, education ana-lyst Henry Braun examined changes in NAEPscores for states with and without high-stakesstandards and tests. He found, as others had,that high-stakes states showed greater grade-level improvements on NAEP than did stateswith lesser stakes. However, when Braunlooked at improvements for specific cohortsof students, such as between 1996 fourth-grade scores and 2000 eighth-grade scores,low-stakes states had an advantage. He alsoreiterated a very important caveat: “conclu-sions from analyses concerning the effects oftesting or, more generally, accountability poli-cies must be tentative, based as they are onhighly aggregated, observational data.”62

In a 2005 article, researchers HendrikJürges, Kerstin Schneider, and Felix Büchelattempted to determine the effect of centralexit exams while better controlling for vari-ables such as culture. They did so through adifference-in-differences framework thatlooked at the differential for vocational-trackstudents between TIMSS mathematics andscience scores in German states with andwithout centralized exit exams. This helpedto estimate the effect of unobserved variablesbecause no states with central examsrequired vocational-track students to takethem in science, while all required mathe-matics. If the exams drove performance, onewould expect to see bigger math-science dif-ferentials in exit exam than non-exam states.

Jürges, Schneider, and Büchel did havesome concerns. First, if improving mathe-

matics skills also improves science outcomes,the effect of exit exams could be understated,with science scores keeping up with math inpart because of the math exams. This did notgreatly concern the authors, however,because only a small fraction of TIMSS sci-ence questions called on math skills. On theflip side, testing math could “crowd out” sci-ence instruction, overstating the effect of exitexams. In an attempt to control for this, theauthors included as an independent variablethe number of hours students reportedlyspent studying science outside of school.Finally, it is important to keep in mind that“college-prep” students were not included inthis study, limiting the applicability of itsfindings to those students.

The authors found that their analysis dra-matically cut the “raw” exit examinationadvantage: from about 1.25 school years toan estimated one-third of a year, bolsteringthe culture theory while suggesting thatexams have some independent positiveeffect.63 The study also provided anotherindication that differing cultures and ideolo-gies might affect where centralized examswould be found and, perhaps, be more effec-tive: except for Bavaria, all of the Germanstates with central exit exams were in the for-mer East Germany.

In the same year that Jürges, Schneider,and Büchel’s work was published, Hanushekand Raymond attempted to identify the effectof imposing state standards and tests by com-paring states’ NAEP score growth fromfourth to eighth grade, which would include asingle cohort of students. Looking at scoresbetween 1992 and 2002, they found that stateaccountability systems had a positive effecton achievement. They also provided potentialinsight into the culture question with theirown difference-in-differences design, whichcontrolled for unobserved state variables“that exert a constant influence on state per-formance over the relevant observation peri-od.”64 This could have controlled for culture,but the period covered—essentially, the early1990s is the entire no-standards period—istoo short to reach any firm conclusions about

14

Except forBavaria, all of the

German stateswith central exit

exams were in theformer East

Germany.

changing culture or attitudes. In addition,although the authors provided an analysiswithout four outliers—California, NorthCarolina, Texas, and Washington, DC—andfound that the outliers had little effect onoverall results, they offered no data brokendown by state showing how individual statesdiffered from overall averages.

The Research Summarized

Having reviewed the extant comparativeresearch on standards set both at national andsubnational levels, it is clear that the empiricalsupport for national standards is weak. First,there simply has not been much comparativeresearch done. Second, what little has beendone has typically focused not simply onnational standards, but standards coupledwith high stakes for students, something notcontemplated, at least publicly, under theCommon Core State Standards Initiative.Finally, the research does not control ade-quately for unobserved variables, especiallyculture or changes in attitudes, that could dri-ve both centralization and a focus on easilytested academic achievement.

Problems beyond theWeak Research Base

The first question that needs to beanswered before fundamentally altering thestatus quo is whether a given reform willwork. The existing national-standards re-search provides no conclusive answer.Beyond the huge deficiencies in the empiricalcase, however, there are other importantobjections to imposing national standardson American education.

Federal and InvoluntaryMany national standards advocates are

quick to point out that the CCSSI—the lead-ing national standards effort—involves statescooperatively determining standards and vol-untarily adopting them, not federal control.

As the CCSSI website explains: “Governorsand state commissioners of education fromacross the country committed to joining astate-led process to develop a common coreof state standards in English-language artsand mathematics for grades K–12.”65

There is a sound political reason foremphasizing state control: citizens and stateofficials are loath to have standards imposedon them by Washington, and previous effortsto create standards at the federal level failedmiserably.66 Moreover, many people fear thathaving the federal government set standardsand write tests would make it easy for special-interest groups like teacher unions andadministrator associations to keep standardslow. They would essentially have “one-stopshopping,” a single power center on which tofocus all of their energies. As Sandy Kress, achief architect of the No Child Left BehindAct, has stated, “the process [of setting stan-dards] will be hijacked by [interest] groups ifthe process is federal.”67

That said, despite national–standardssupporters emphasizing that the CCSSI isstate-led and that adoption of its standards istechnically voluntary, adoption will almostcertainly be de facto involuntary, and thestandards themselves ultimately federal.Already, as previously noted, Secretary ofEducation Duncan has made clear that itwould behoove states to sign on to the CCSSIif they wish to compete for a share of the$4.35 billion “Race-to-the-Top” fund.Duncan has also said that the federal govern-ment would furnish $350 million to developtests tied to the standards. Finally, it wouldbe very difficult to conceive of a reauthorizedNo Child Left Behind Act, especially in lightof what Duncan is currently pushing, thatwould not mandate adoption of nationalstandards and ultimately lead to federal con-trol of those standards. Of course, statescould “voluntarily” turn down the billions offederal dollars attached to NCLB, but even inthat unlikely case there would still beabsolute compulsion involved: state citizenswould continue to have no choice about pay-ing federal taxes.

15

Beyond the hugedeficiencies in theempirical case,there are otherimportant objections toimposing nationalstandards.

ConstitutionalityThe federal government has only been

heavily involved in education since 1965,when the Elementary and SecondaryEducation Act was passed, because for most ofthe nation’s history it was understood thatfederal involvement would be unconstitution-al. The Constitution makes no mention of“education” or “schooling” among the specif-ic, enumerated powers it gives to the federalgovernment, and outside of controlling theDistrict of Columbia and military educationalactivities, Washington has no authority to beinvolved in education.68

To justify its growing involvement, federalpolicymakers have typically argued thatWashington does not force states and schoolsto do anything, but only attaches rules andregulations to federal money that states anddistricts may turn down. By essentiallydemanding that all states and districts adoptspecific standards, however, Washingtonwould be exceeding even the unconstitution-al power it has accumulated under ESEA, vio-lating a stipulation that has been in federaleducation law—including NCLB—since dayone: “Nothing in this Act shall be construedto authorize an officer or employee of theFederal Government to mandate, direct, orcontrol a State, local education agency, orschool’s curriculum, program of instruction,or allocation of State or local resources.”69

MeasurementThere is a great deal of debate within the

education community over the importanceof “hard” learning, such as memorizing factsor being able to complete mathematicaloperations, and “soft” learning of such skillsas creativity and critical thinking. Most stan-dardized tests, by their very nature, primarilymeasure the former. But there is no convinc-ing evidence that such hard learning is neces-sarily the most important outcome of aneducation system.

Given that all people are unique, the goalsthat they have for their education and that oftheir children is potentially as varied as thepeople themselves. Assume, though, that the

end goal for all people is happiness, whetherachieved through a well-paying job, doingsomething that is intellectually fulfilling, orjust sitting on a couch watching television allday. Taking happiness maximization as a rea-sonable final aim for education, it appears atleast on the surface that international acade-mic assessments are poor measures ofwhether or not people are getting what theywant out of education. Of the 10 top-finish-ing countries on the 2007 TIMSS eighth-grade mathematics exam, only four—SouthKorea, Singapore, Japan, and the UnitedStates—ranked among the top 50 in “subjec-tive well-being,” and among those the UnitedStates ranked the highest by far on happi-ness.70 Of course, a lot beyond educationgoes into happiness, and measuring that feel-ing is a tricky and imperfect undertaking, butthese results at least plausibly suggest that aneducation system focused on achievementdemonstrable by standardized test scoresmight not be optimal.

Coping with DiversityA major roadblock for establishing cen-

tralized standards in any political system oth-er than a dictatorship is getting enough peo-ple to agree on their substance. What shouldthey include? What shouldn’t they include?Who should write them? This becomesincreasingly difficult the more linguistically,religiously, ideologically, and ethnicallydiverse a nation’s population. With that inmind, it should be no surprise that many ofthe nations that have national standards andtend to perform best on international examsalso tend to be highly homogeneous.

Consider Finland, the top-place finisher onthe 2006 PISA assessment. With only about 5million people, it has a population less than 2percent that of the United States. In terms ofmobility, it has net migration of 0.68 immi-grants per 1,000 people, versus 4.31 per 1,000Americans. It only has two major ethnicgroups—Finns are roughly 93 percent of thepopulation and ethnic Swedes about 6 per-cent—while the United States has a populationthat is predominantly “white” but 13 percent

16

By demandingthat all states

adopt specificstandards,

Washingtonwould be

exceeding even theunconstitutional

power it has accumulatedunder ESEA.

African-American, 15 percent Hispanic, and 4percent Asian. Religiously, about 83 percent ofFinns belong to the Lutheran Church ofFinland and most of the rest profess no reli-gion. In contrast, about 51 percent ofAmericans are Protestant—which includesmany denominations—24 percent RomanCatholic, 12 percent unaffiliated, 1.7 percentJewish, and the rest members of several smallerreligious groups. Finally, roughly 91 percent ofFinns speak Finnish, 6 percent speak Swedish,and there are small Sami- and Russian-speak-ing minorities. In the United States, about 82percent of people speak English, 11 percentSpanish, and the rest many other languages.71

How about South Korea, another regulartop-place finisher? Its total population is 48.5million people, about 16 percent the size ofthe United States. Rather than gainingmigrants, South Korea is losing them.Ethnically, South Korea is almost completelyhomogeneous. Linguistically, everyone speaksKorean. The only way in which South Korea isremotely similar to the United States is in reli-gious diversity, with about one-quarter of itspeople Christians (with significant numbersof Protestants and Roman Catholics amongthem), about 23 percent Buddhists, and halfclaiming no religious affiliation.72

Finally, consider Germany, which althoughnot a high-flyer on international assessments,is the nation that William Schmidt points toas proof that the United States can create andadopt coherent, rigorous national standards.

With about 82 million people, Germany’spopulation is roughly a quarter the size ofthe United States’ but bigger than SouthKorea’s and much larger than Finland’s. Ithas 2.19 immigrants per 1,000 people, anumber about half that of the United Statesbut much larger than Korea and Finland.Ethnically, about 92 percent of Germany’spopulation is German, 2.4 percent Turkish,and the rest a smattering of other groups.This could be considered on par withFinland, although the gulf between ethnicFinns and Swedes is not as large as betweenGermans and Turks. The population of theUnited States, however, is much more diverse

than that of Germany. Religiously, aboutone-third of Germans are Protestants, one-third Roman Catholics, 3.7 percent Muslim,and the rest unaffiliated—diversity that isclose to the United States and much greaterthan Finland. Finally, the only native lan-guage spoken in Germany is German, anoth-er significant departure from the U.S. experi-ence but also less diverse than Finland.73

Clearly, two of the top performers oninternational assessments are very homoge-neous, and Germany—which performsroughly on par with the United States and ismoving toward national standards—is morediverse than the top performers but appre-ciably more homogeneous than the UnitedStates. This suggests that diversity militatesagainst establishing national standards and,perhaps, high performance on standardizedtests. Because people of different back-grounds often have different priorities anddemands for education, it is logical to expectthat the greater the diversity of the peoplefalling under a single schooling authority,the greater the conflict, the less coherent thecurriculum, and the worse the outcomes.

Beyond just logic, while there is no system-atic empirical support for the assertion above,some broad evidence is highly suggestive.Political scientist Robert Putnam has foundthat greater social capital—essentially, thebonds of trust between people—is correlatedwith better academic outcomes, and thatgreater social capital is found in areas with lessdiversity.74 Similarly, sociologist JamesColeman has reported that after controllingfor variables such as students’ socioeconomicstatus, Roman Catholic schools have betteracademic outcomes because they are “closed”systems in which all actors—students, par-ents, and school employees—share importantnorms and values.75

What Can Be Done?

Culture, politics, diversity—all of theseforces militate against the success of nation-al standards in the United States. But if not

17

Culture, politics,diversity—all ofthese forces militate againstthe success ofnational standards.

national standards, what is the solution toour educational woes? We have tried localcontrol of public schools. We have tried statecontrol. What else can be done?

The question, ultimately, is what type ofreform can both successfully navigate thecultural, political, and diversity shoals onwhich so many public-schooling reformshave run aground and deliver optimal educa-tional outcomes? The answer is to replacepublic schooling—in which government notonly ensures that all children can access edu-cation, but also provides the schools—withtrue public education. Let parents control edu-cation funds either through vouchers or,preferably, personal and donation tax credits,and let them freely choose among autono-mous schools and other educational opt-ions.76 Let education work as a free market, inwhich consumers purchase services andproducts according to their individual needsand desires, and suppliers compete throughquality, specialization, price, and innovation.

The Empirical Support for EducationalFreedom

The first argument for educational free-dom is that, unlike national standards, theempirical evidence supporting it is broad andconvincing.

For one thing, there is a wealth of evidenceshowing that free markets, overall, meet theneeds and desires of people much better thancommand economies, which is essentiallywhat public schooling is.77 Free markets fos-ter competition that drives innovation andimprovement, all while enabling consumerswith unique needs to seek out producers thatare willing and able to meet those needs.

In education, the research is also deep andconvincing; including evidence based on stan-dardized achievement tests. Ten prominentempirical studies have analyzed voucher pro-grams using the “gold-standard” social sciencetechnique of randomly assigning subjects to“treatment” or “control” groups. Nine ofthose studies found that at least one subgroupof students receiving vouchers did better onacademic assessments than those who applied

for but did not receive vouchers, while in nosubgroup did voucher students do worse. Inthe tenth study, the results were a wash.78

Comparing the outcomes of educationsystems around the world based on howfreely consumers and providers can interactalso bears this out. Reviewing 65 studies thatreported over 156 separate statistical find-ings, Andrew Coulson found that findingsfavoring free-market provision of educationoutnumbered those favoring government-monopoly provision by a ratio of 15-to-1.79

Notably, virtually all of these studies com-pared school systems in close proximity toone another within individual nations, whichwould mitigate the potentially confoundingeffects of cultural and other difficult-to-observe variables. Similarly, education profes-sor James Tooley has found low-cost private(often for-profit) schools to be ubiquitous inmany of the world’s poorest slums, and thatthey regularly outperform “free” governmentschools on assessments of such things asnative-language mastery, English, and math-ematics.80

Going Beyond What Is Testable So choice produces better results on

achievement assessments. However, allowingpeople not to fixate exclusively on easily test-ed outcomes might also enable them to focuson other pursuits or skills best suited to theirabilities and ultimate desires. As already dis-cussed, Americans typically express a greaterdegree of happiness than their peers innations with more centralized, test-driveneducation systems. If enabling people tomaximize their happiness is the end goal ofeducation—and anything else we do, for thatmatter—then allowing students to freely pur-sue their own educational interests (with theguidance of parents or other advisers) wouldlikely produce the best educational out-comes, if not necessarily the best test scores.

Recent actions in countries that often doquite well on international examinations—nations that many national–standards advo-cates hope to emulate—suggest that they are,perhaps, coming to believe that education

18

The first argument for

educational freedom is that

the evidence supporting it is

broad and convincing.

cannot and should not be reduced simply totest scores. For instance, although Japan hasrecently retrenched somewhat, as early as the1970s its education ministry started encour-aging schools to put more emphasis on suchsoft skills as “critical thinking.” In 2002,Japan shrank its school week from six days tofive, and reduced the content of its nationalcurriculum by about 30 percent. It also intro-duced a class called “Integrated Studies” thatfeatures more independent, student-drivenwork.81

Starting in 1999, Singapore, another topperformer, reduced its national curriculumby about a third, doing so in order to “pro-vide room for teachers to implement the keyinitiatives announced in 1997, namely theinfusing of thinking skills and integratingthe use of Information Technology inlessons.”82 In 2001, Singapore further re-formed its curriculum, focusing it even moreon teaching students to “think creatively andapply knowledge innovatively.” It also intro-duced ability-based tracking.83 Finally, fund-ed in part by Korean education ministries,Korean teachers have been regularly comingto the United States to learn how to teachcreativity and critical thinking in theirschools.84

Does this mean that easily tested subjectshave been deemed irrelevant in these high-flying nations? And couldn’t administrators,teachers, and other interest groups just betrying to make their lives easier with thesereforms, just as they do in the United States?

Movement away from strict standardsand testing does not prove anything, ofcourse. These nations probably still stronglysupport centralized standards and testing,and parents might vehemently object to soft-ening curricula. Still, that these nations withlong, proven track records of supportingtough centralized standards and testing havemoved to loosen their systems must not beignored—it could mean that they have recog-nized real shortcomings of their systems.

Japan, for instance, suffered the “lostdecade” of the 1990s: a long period of eco-nomic malaise that many experts have at least

partially attributed to the nation’s lack of cre-ative thinking and hidebound social struc-tures.85 Bolstering this is the fact that Japanhas not been outpacing the United States eco-nomically, its significantly higher test scoresnotwithstanding. In 1980, Japanese GDP percapita, adjusted for purchasing power, was$8,901. In the United States it was $12,255.Since then, per capita Japanese GDP hasgrown about $855 per year, versus almost$1,180 in the United States.86

Minimizing Poisonous PoliticsIn addition to enabling individuals to

pursue educational options best suited totheir desires and needs, free-market educa-tion avoids the political fighting that hasoften doomed standards-based reforms. Bytaking control of education away from gov-ernment, much of the politics that handicapsthe delivery of education would automatical-ly be removed.

Currently, because all taxpayers have tosupport government schools, almost everyschool or district decision is politicized.District budgets constantly spark conflict.There are incessant curricular and pedagogi-cal battles over everything from the grade atwhich students start to use calculators towhat reading programs schools employ.87

And of course there is the asymmetrical pow-er problem, in which the people employed bythe public schools have disproportionatepower compared to those people whom theschools are supposed to serve.

Give parents control over education dol-lars and let them choose among autonomousschools, and these problems would diminishconsiderably. District budgets would be irrel-evant because schools would be fundedbased on their ability to attract customers,not their allocations of public dollars.Curricular and pedagogical battles wouldpeter out as schools independently chose thecurricula they thought best and parentsselected the schools with programs theythought most effective for their children.Finally, no longer would going through alabyrinthine and stacked political process be

19

Free-market education avoidsthe politicalfighting that hasdoomed standards-basedreforms.

the only avenue through which “customers”could try to fix education problems; theywould be able to execute immediate account-ability by taking their children and the mon-ey to educate them out of unsatisfactoryschools and putting them elsewhere.

Even with full choice, it should be noted,there would still be politics in education. Forone thing, there would no doubt always bepeople calling for greater or less regulation ofschools. In addition, what the minimum edu-cational requirements should be for all chil-dren would likely be a contested question.Finally, depending on the mechanism usedto deliver school choice, there would nodoubt be recurring disputes about the maxi-mum voucher or tax credit size, as well as eli-gibility criteria.

Even with those lingering problems, how-ever, the amount of education subject to pub-lic-policy decisions would be much smallerthan under the status quo. And as AndrewCoulson has empirically demonstrated, ifchoice is delivered through tax credits ratherthan vouchers—meaning no tax dollars gothrough government and the money headsto recipients according to the free choice ofthe sender—onerous regulation is much lesslikely.88

Dealing with DiversityFinally, the last great stumbling block that

makes improving government schooling verydifficult is diversity—ethnic, religious, andideological. It has led to the neutering of pub-lic school curricula and the removal of poten-tially contentious—but also important andinteresting—content from the nation’sschools. As Ravitch has documented, drives toeliminate objectionable content from curricu-la have come from all parts of the social andpolitical spectra, with conservatives generallyattacking morally objectionable, or insuffi-ciently “American,” content and liberalsdecrying anything that could be consideredcritical or insensitive to various minoritygroups. This has resulted in textbooks thatare devoid of engaging content and curriculadesigned to skirt controversy.89 And this cur-

ricular denuding is not just demonstrated inovert actions. Based on widespread anecdotalevidence, it appears that to avoid conflictmany teachers skip over controversial topicssuch as evolution even when state standardsrequire that they be taught.90 Perhaps that iswhy, despite religious instruction having longbeen banished from public schools, close tohalf of all Americans believe that humanbeings and other living creatures were created,most likely by God, in their present forms.91

How would school choice negotiate thesetreacherous shoals? In the same way it wouldavoid so many political fights: by letting peo-ple select educational options compatiblewith their diverse norms and backgroundsrather than requiring them to fight for con-trol of a single system. Parents who wanttheir children to learn religious explanationsfor human origins would be able to patronizelike-minded schools, as would the strictestatheists. Hispanic families that desire in-struction in Hispanic history and languagecould seek schools that teach them. Libertar-ians who want their children taught that theNew Deal prolonged the Great Depressioncould look for schools that taught that.Then, rather than militating against any andall children hearing things that are potential-ly controversial, diversity would be fullyaccommodated and children would be ableto get coherent instruction.

With Choice, Would There Be Standards?One of the most common objections to

choice—and one that is especially germane tothe topic of national standards—is that free-market delivery of education would result inan absence of standards. It would create astate of relativistic, educational anarchy,according to critics.

It is true that choice would potentially let“a thousand flowers bloom” and that ideasand knowledge that might turn out to beincorrect would and could be taught. Thislatter possibility is something that national–standards advocates implicitly assert thatthey could prevent by choosing the “best”curriculum for everyone. But no one person

20

No one person orgroup has a

monopoly ontruth or

understanding.

or group of people has anything close to amonopoly on truth or understanding, andthe best system to deal with such imperfec-tion is one in which power is diffuse, not con-centrated. It allows ideas to compete, andthose that turn out to be wrong would nothave been imposed on everyone.

That said, it is inaccurate to assert thatwidespread standards could not and wouldnot be adopted in a laissez-faire educationsystem. Consider the long period of Ameri-can history—from the arrival of the firstBritish colonists to the common-schoolsmovement in the mid-1830s—during whichmost education was delivered in homes, bychurches, and in other nongovernmentalarrangements. During that time the very“American” language that we speak was stan-dardized by just a few commercially avail-able tools. Noah Webster was explicit in hisaims to create a standard American languageand forge a unique American culture withhis American Spelling Book and AmericanDictionary of the English Language, and heachieved great success in that endeavor.Indeed, he had sold 20 million copies of hisspeller by 1829, although by 1830 the totalU.S. population was less than 13 million.Several other people also wrote and soldreaders during this time, and the renownedand ubiquitous McGuffey’s Reader was pub-lished in 1836.92

Today we have countless examples of non-governmental standardization both insideand outside of education. Millions of stu-dents around the country take the SAT orACT, as well as Advanced Placement orInternational Baccalaureate tests, for themost part by choice. Similarly, U.S. News andWorld Report, Forbes, the Princeton Review,and numerous other organizations provideguides and rankings to colleges, which unlikeK–12 schools, are chosen by students.Outside of education we have UnderwritersLaboratory, Consumer Reports, franchising,and sundry other mechanisms setting stan-dards for innumerable items and services weuse every day. Standards are ubiquitous infree markets.

Conclusion

The argument on behalf of national educa-tion standards is alluringly simple: set highstandards that all American schools and stu-dents must meet, and all students will achieveat high levels. But actually setting high stan-dards and getting students to meet them isextremely difficult. The opposition of the mostpolitically powerful interests in educationmust be overcome. Crippling conflicts betweengroups with different religious, ethnic, and ide-ological concerns must be avoided. And finally,a culture that is averse to intense focus on aca-demics and the kind of “hard,” largely mathe-matical and scientific knowledge that is easy tocapture on a test must be transformed.

Much of this explains why what littlecomparative research there is on nationalstandards is inconclusive. Nations varymarkedly not just in terms of economics andpolitics, but culture and other variables thatare extremely difficult, if not impossible, tocleanly control for in social research. Thesame applies to subnational political units,such as U.S. states and Canadian provinces.And it is hardly a settled question that testscores such as those used in standardsresearch even encompass all the most desirededucational outcomes.

For all of these reasons, the road to suc-cessful education reform appears to go in theopposite direction of greater top-down con-trol. The key appears to be to give educationfunding to parents, allow schools autonomy,and as a result make schools respond to theneeds and demands of parents and children.That would solve the asymmetrical powerproblem, forcing educators to satisfy cus-tomers rather than use politics to get theirway. It would prevent paralyzing political,cultural, religious, or ethnic conflicts thatforce lowest-common-denominator stan-dards. And it would lead to standards thatwould be meaningful, but also sufficientlyflexible so that unproven ideas could com-pete, and inevitable human failures wouldnot be inflicted on everyone. In other words,rather than centralizing crippling govern-

21

Standards areubiquitous in free markets.

ment power even further, it would trulyreform the education system.

Notes1. U.S. Department of Education, A Nation at Risk,archived information, April 1983, www.ed.gov/pubs/NatAtRisk/risk.html.

2. Among several major reports calling for na-tional standards over the last few years are ChesterE. Finn Jr., Liam Julian, and Michael Petrilli, ToDream the Impossible Dream: Four Approaches toNational Standards and Tests for America’s Schools(Washington: Thomas B. Fordham Foundation,2006); The Commission on No Child Left Behind,Beyond NCLB: Fulfilling the Promise to Our Nation’sChildren (Washington: The Aspen Institute, 2007),p. 127; and National Governors Association,Council of Chief State School Officers, andAchieve, Inc., Benchmarking for Success: Ensuring U.S.Students Receive a World-Class Education (Washing-ton: National Governors Association, 2008).Among recent high-profile individual endorsersof national standards have been former U.S. sec-retaries of education Rod Paige and William J.Bennett, “Why We Need a National School Test,”Washington Post, September 21, 2006; andAmerican Federation of Teachers president RandiWeingarten, “The Case for National Standards,”Washington Post, February 16, 2009.

3. Council of Chief State School Officers andNational Governors Association, “Common CoreState Standards K–12 Work and Feedback GroupsAnnounced” press release, November 10, 2009.

4. Alyson Kline, “To Duncan, Incentives a Priority,”Education Week, January 30, 2009, www.edweek.org/ew/articles/2009/02/04/20duncan.h28.html?r=2063662804; U.S. Department of Education,“Duncan Offers Stimulus Funds for States toDevelop Rigorous Assessments Linked to Com-mon Standards” (press release, June 15, 2009).

5. Constitutionally, it is clear that states haveauthority over education. Article I, Section 8, of theU.S. Constitution gives the federal government nopower to govern American education, and the 10thamendment reinforces that in stating that “thepowers not delegated to the United States by theConstitution, nor prohibited by it to the States, arereserved to the States respectively, or to the people.”In addition, even analysts who reject this interpre-tation of the Constitution typically acknowledgethat education has traditionally been the domainof state governments, not Washington.

6. Finn et al., p. 11–12.

7. Victor Bandeira de Mello, Charles Blankenship,and Don McLaughlin, Mapping State ProficiencyStandards Onto NAEP Scales: 2005–2007 (Washing-ton: National Center for Education Statistics,October 2009).

8. Duncan quoted in Oliver Staley and MollyPeterson, “Duncan to Spend Billions to ‘Trans-form’ U.S. Schools,” Bloomberg.com, April 16,2009, www.bloomberg.com/apps/news?pid=20601103&sid=aXp_0t1dgo_Q&refer=us.

9. National Governors Association, Council ofChief State School Officers, and Achieve, Inc., p. 7.

10. Robert M. Costrell, “A Simple Model of Edu-cational Standards,” American Economic Review 84,no. 4 (1994): 957–64.

11. Robert M. Costrell, “Can Centralized Educa-tional Standards Raise Welfare?” Journal of PublicEconomics 65 (1997): 271–93.

12. Conference Board, Corporate Voices forWorking Families, Partnership for 21st CenturySkills, and Society for Human Resource Manage-ment, Are They Really Ready to Work: Employers’Perspectives on the Basic Knowledge and Applied Skillsof New Entrants to the 21st Century U.S. Workforce,2006, www.21stcenturyskills.org/documents/FINAL_REPORT_PDF09-29-06.pdf.

13. Andrew Flagel, “Admission to College: Does MyHigh School Matter?,” Not Your Average Admis-sions Blog, notjustadmissions.com/2007/11/01/admission-to-college-does-my-high-school-matter/,November 1, 2007.

14. Diane Ravitch, National Standards in AmericanEducation: A Citizen’s Guide (Washington: Brook-ings Institution Press, 1995), pp. 25–27.

15. Ibid., pp. xxiii–xxiv.

16. NEA and AFT headquarters employment countderived from the U.S. Department of Labor’s union“Officer/Employee Search” database at kcerds.dol-esa.gov/query/getOfficerEmployeeQry.do. Search-ing under the name of the union and “nationalheadquarters” generates an employment report foreach “employee” or “officer” employed by theunion during a given fiscal year. This likely over-counts the total number of employees at any giventime during the year due to normal employeeturnover, but it nonetheless provides a generalsense of both unions’ national presence.

17. For an excellent discussion of the history of thePTA and de facto NEA control over the group, seeCharlene K. Haar, The Politics of the PTA (New Bruns-wick, NJ: Transaction Publishers, 2002), pp. 67–84.

22

18. LM-2s are available at the U.S. Department ofLabor, “Officer/Employee Search” database,kcerds.dol-esa.gov/query/getOfficerEmployeeQry.do.

19. Ballot Initiative Strategy Center, “What WeDo,” www.ballot.org/pages/what_we_do.

20. Economic Policy Institute, “A Broader, BolderApproach to Education” (report summary, June25, 2009), www.boldapproach.org/report_20090625.html.

21. Fairtest, “The Case against High StakesTesting,” www.fairtest.org/k-12/high+ stakes, andstatements from unions in California,Massachusetts, Missouri, Nebraska, and Wash-ington State at www.fairtest.org/Organizations+and+Experts+Opposed+to+High+Stakes+ Testing.

22. Paul Houston, American Association of SchoolAdministrators, “High-Stake Testing Overwhelm-ingly Gets an ‘F’” press release, June 19, 2000.

23. Vincent L. Ferrandino and Gerald N. Tirozzi,“Principals’ Perspective: An Open Letter to the110th Congress,” www.principals.org/s_nasspbin.asp?CID=1237&DID=55577&DOC=FILE. PDF.

24. National Council of Teachers of English,“Resolution on Urging Reconsideration of HighStakes Testing” (position statement), www.ncte.org/positions/statements/highstakestestrecons.

25. Chester E. Finn Jr., Liam Julian, and MichaelPetrilli, The State of State Standards 2006 (Washing-ton: Thomas B. Fordham Foundation, 2006), p. 16.

26. On the last point, national–standards sup-porters will argue that polls show that Americanssupport standardized testing and even a nationaltest. See, for instance, Phi Delta Kappa/Gallup,“Highlights of the 2009 Phi Delta Kappa/GallupPoll,” Phi Delta Kappan, September 2009, p. 8.However, while respondents might like commonstandards and testing in the abstract, it is entirelypossible they don’t like it when it becomes reality.Hence, there is high approval for “accountability,”but low approval for No Child Left Behind(NCLB). For a good discussion of this, see DavidJ. Hoff, “Polls find NCLB is Unpopular, ButAccountability Isn’t,” NCLB: Act II, August 22,2008.blogs.edweek.org/edweek/NCLB-ActII/2008/08/dissecting_polls_nclb_ is_unpop.html.

27. Seymour Martin Lipset, “Still the ExceptionalNation?” Wilson Quarterly 24, no. 1 (Winter 2000).

28. J. M. Roberts, A Short History of the World (NewYork: Oxford University Press, 1993), pp. 196 and206. International development scholar FrancisFukuyama notes that though the Japanese have

typically been more centered on sub-state-levelgroups than the Chinese, they too have a very col-lective-oriented culture, as well as a culture of“lifetime employment” with one company thatdrives fixation on doing well on college entranceexams and getting into a well-regarded school.Francis Fukuyama, Trust (New York: Free Press,1995), pp. 54 and 186–88.

29. Tom Loveless, The 2001 Brown Center Report onAmerican Education 1, no. 2 (September 2001):18–28, and The 2002 Brown Center Report onAmerican Education 1, no. 3 (September 2002):16–28.

30. Harold W. Stevenson, Chuansheng Chen, andShin-Ying Lee, “Mathematics Achievement ofChinese, Japanese, and American Children,”Science 259, no. 5091 (January 1, 1993): 53–58.

31. For an in-depth discussion of the impact ofculture on academic outcomes, including, specif-ically, Asian students, see Abigail and StephanThernstrom, No Excuses: Closing the Racial Gap inLearning (New York: Simon and Schuster, 2003).

32. Weingarten.

33. Trends in International Mathematics andScience Study (TIMMS) mathematics scores fromInstitute for Education Sciences, “Trends inInternational Mathematics and Science Study:Table 1. Average mathematics scores of fourth-and eighth-students, by country: 2007,” nces.ed.gov/timss/table07_1.asp, and “Trends in Interna-tional Mathematics and Science Study: Table 3.Average science scores of fourth- and eighth-stu-dents, by country: 2007,” nces.ed. gov/timss/table07_3.asp. Nations with and without nationalstandards are identified in V. S. Mullis et al., eds.,TIMSS 2007 Encyclopedia: A Guide to Mathematicsand Science Education Around the World, Volume 1 A-L, TIMSS and Progress in International ReadingLiteracy Study (PIRLS) International StudyCenter, Exhibits 5 and 6, 2008, timss.bc.edu/TIMSS2007/encyclopedia. html. Note that theEncyclopedia identifies the presence of national“curriculum.” For consistency, and since the term“national standard” is somewhat vague when usedin policy discussions, the term “standard” hasbeen substituted for “curriculum” in this paper.

34. Scores from TIMSS International StudyCenter, “TIMSS Highlights: February 1998,”timss.bc.edu/timss1995i/TIMSSPDF/C_Hilite.pdf.Level of curricular decisionmaking from Ina V.S .Mullis et. al., Mathematics and Science Achievementin the Final Year of Secondary School: IEA’s Third Inter-national Mathematics and Science Study (ChestnutHill, MA: Boston College, 1998), p. 25.

35. Scores from Organization for Economic

23

Cooperation and Development, PISA 2006 ScienceCompetencies for Tomorrow’s World, ExecutiveSummary, p. 22, www.oecd.org/dataoecd/15/13/39725224.pdf. Standards-based external exami-nations data are from PISA 2006 Science Competen-cies for Tomorrow’s World, Table 5.2, www.oecd.org/dataoecd/30/63/39704360.xls.

36. PISA 2006 Science Competencies for Tomorrow’sWorld, Table 5.22.

37. John H. Bishop, “The Impact of Curriculum-Based External Examinations on School Prioritiesand Student Learning,” International Journal ofEducational Research 23, no. 8 (1996): 653–752. In“The Effect of National Standards and Curriculum-Based Exams on Achievement,” The AmericanEconomic Review 87, no. 2 (1997): 260, Bishop refersto “Curriculum-Based External Exit Examinations,”but both that and “Curriculum-Based ExternalExaminations” refer to standards-and-testingregimes with high-stakes consequences for sec-ondary students.

38. Bishop, “The Impact of Curriculum-BasedExternal Examinations,” pp. 698–99.

39. Ibid., p. 697.

40. Ibid., pp. 697–99.

41. Ibid., p. 699.

42. Ibid., p. 260.

43. John H. Bishop, “Are National Exit Examina-tions Important for Educational Efficiency?”Swedish Economic Policy Review 6 (1999): 349–98.

44. Ibid., pp. 374–75.

45. Ludger Woessmann, “Central Exit Exams andStudent Achievement: International Evidence,” inNo Child Left Behind: The Politics and Practice of SchoolAccountability, ed. Paul E. Peterson and Martin R.West (Washington: Brookings Institution Press,2003), pp. 292–323. Woessmann performed a sim-ilar analysis in “Schooling Resources, EducationalInstitutions and Student Performance: the Inter-national Evidence,” Oxford Bulletin of Economics andStatistics 65, no. 2 (2003). He looked at the presenceof centralized exams as one among many institu-tional and resource variables and reached similarconclusions.

46. Woessmann, p. 304.

47. Assignment of countries to regions providedin an e-mail to the author by Ludger Woessmann.

48. Ludger Woessmann, “How Central ExamsAffect Educational Achievement: International

Evidence from TIMSS and TIMSS-Repeat,” publi-cation number PEPG/02-10 (Cambridge, MA:Program on Education Policy and Governance,Harvard University, 2002).

49. Hendrik Jurges and Kerstin Schneider,“International Differences in Student Achieve-ment: An Economic Perspective,” GermanEconomic Review 5, no. 3 (2004): 371.

50. William H. Schmidt, Richard Houang, andSharif Shakrani, International Lessons about NationalStandards (Washington: Thomas B. FordhamInstitute, 2009), p. 29.

51. William H. Schmidt, “Panel 1: NationalStandards in Other Nations,” (remarks atFordham Foundation conference on nationalstandards, May 5, 2009), www.edexcellence.net/detail/event.cfm?event_id=4&id=317.

52. Organization for Economic Cooperation andDevelopment, “The Programme for InternationalStudent Assessment (PISA),” 2007, www.oecd.org/dataoecd/15/13/39725224.pdf.

53. In “How Well are American Students Learn-ing?” The 2008 Brown Center Report on AmericanEducation (Washington: Brookings Institute, 2009),p. 16, Tom Loveless offers important insight, forinstance, into the Program for InternationalStudent Assessment (PISA) science literacy test. Hereports that it does not just test knowledge, but“considers students’ attitudes, beliefs, and valuespart of literacy. Questionnaires are administered tostudents and some attitudinal items are inter-spersed in the portion of the tests assessing knowl-edge.” This demonstrates the importance of notassuming that results on any given test reflectdesired educational outcomes, and substantiatesconcerns that testing can become politicized—anespecially dangerous problem were a single test tobe imposed on the entire nation.

54. Bishop, “The Impact of Curriculum-BasedExternal Examinations,” pp. 711–24.

55. Audrey L. Amrein and David C. Berliner, “TheImpact of High-Stakes Tests on Student AcademicPerformance: An Analysis of NAEP Results inStates with High-Stakes Tests and ACT, SAT, andAP Test Results in States with High SchoolGraduation Exams,” publication number EPSL-0211-126-EPRU (Tempe, AZ: Education PolicyStudies Laboratory, Arizona State University,December 2002). An earlier version of the study is“High-Stakes Testing, Uncertainty, and StudentLearning,” Education Policy Analysis Archives 10, no.18 (March 28, 2002), epaa.asu. edu/epaa/ v10no18.

56. Barak Rosenshine, “High-Stakes Testing:Another Analysis,” Education Policy Analysis Ar-

24

chives 11, no. 24 (August 4, 2003), epaa.asu.edu/epaa/v11n24/.

57. Audrey Amrein-Beardsley and David C.Berliner, “Re-analysis of NAEP Math and ReadingScores in States with and without High-StakesTests: Response to Rosenshine,” Education PolicyAnalysis Archives 11, no. 25 (August 4, 2003), http://epaa.asu.edu/epaa/v11n25/.

58. Margaret E. Raymond and Eric A. Hanushek,“High-Stakes Research,” Education Next, Summer2003, pp. 48–55.

59. Martin Carnoy and Susanna Loeb, “DoesExternal Accountability Affect Student Outcomes?A Cross-State Analysis,” Educational Evaluation andPolicy Analysis 24, no. 4 (Winter 2002): 313.

60. Carnoy and Loeb, p. 320.

61. Ibid., pp. 314–15.

62. Henry Braun, “Reconsidering the Impact ofHigh-stakes Testing” Education Policy AnalysisArchives 12, no. 1 (January 5, 2004), epaa.asu.edu/epaa/v12n1/.

63. Hendrik Jürges, Kerstin Schneider, and FelixBüchel, “The Effect of Central Exit Examinationson Student Achievement: Quasi-ExperimentalEvidence from TIMSS Germany,” Journal of theEuropean Economic Association 3, no. 5 (September2005): 1153–54.

64. Eric A. Hanushek and Margaret E. Raymond,“Does School Accountability Lead to ImprovedStudent Performance?” Journal of Policy Analysisand Management 24, no. 2 (2005): 302.

65. Common Core State Standards Initiative,www.corestandards.org/.

66. Diane Ravitch describes well how federalefforts to create national standards in the 1990sdegenerated into nationwide acrimony in LeftBack: A Century of Battles Over School Reforms (NewYork: Touchstone, 2000), pp. 429–37.

67. Kress quoted in Finn, Petrilli, and Julian, ToDream the Impossible Dream: Four Approaches toNational Standards and Tests for America’s Schools, p. 20.

68. For more on the constitutionality question,see Neal McCluskey, Feds in the Classroom: How BigGovernment Corrupts, Cripples, and CompromisesAmerican Education (Lanham, MD: Rowman andLittlefield, 2007), pp. 129–64.

69. U.S. Public Law 107-110, 107th Cong., 1st sess.,January 8, 2002.

70. Subjective well-being ranking is from WillWilkinson, “In Pursuit of Happiness Research: IsIt Reliable? What Does It Imply for Policy?” CatoPolicy Analysis no. 590 (April 11, 2007), p. 19.

71. Central Intelligence Agency, “Finland,” TheWorld Factbook, www.cia.gov/library/publications/the-world-factbook/geos/fi.html, and “The UnitedStates,” www.cia.gov/library/publications/the-world-factbook/geos/us.html.

72. Ibid., “Korea, South,” The World Factbook,www.cia.gov/library/publications/the-world-fact-book/geos/ks.html.

73. Ibid., “Germany,” The World Factbook, www.cia.gov/library/publications/the-world-factbook/geos/gm.html.

74. Findings correlating social capital with academicoutcomes are from Robert D. Putnam, BowlingAlone: The Collapse and Revival of American Community(New York: Simon and Schuster, 2000), pp. 299–306;and correlation of social capital with diversity isfrom Putnam, “E Pluribus Unum: Diversity andCommunity in the Twenty-first Century,” Scandi-navian Political Studies 30, no.2 (June 15, 2007): 149.

75. James S. Coleman, “The Creation and De-struction of Social Capital: Implications for theLaw,” Notre Dame Journal of Law, Ethics, and PublicPolicy 375 (1987–1988): 375–404.

76. For an excellent explanation of how and whyeducation tax credits work, see Adam Schaeffer,“The Public Education Tax Credit,” Cato PolicyAnalysis no. 605 (December 5, 2007).

77. For a powerful discussion of the overwhelm-ing evidence of this, see Johan Norberg, In Defenseof Global Capitalism (Washington: Cato Institute,2003).

78. For a review of those studies with links to attainthem, see Jay P. Greene, “Voucher Effects onParticipants,” Jay P. Greene’s Blog, jaypgreene.com/2008/08/21/voucher-effects-on-participants/.

79. Andrew J. Coulson, “Comparing Public,Private, and Market Schools: The InternationalEvidence,” Journal of School Choice 3 (2009): 31–54.

80. James Tooley, The Beautiful Tree: A PersonalJourney Into How the World’s Poorest People AreEducating Themselves (Washington: Cato Institute,2009), pp. 178–82.

81. Christopher Bjork and Ryoko Tsuneyoshi,“Education Reform in Japan: Competing Visionsfor the Future,” Phi Delta Kappan 86, no. 8 (April2005): 620–21.

25

82. Singapore Ministry of Education, “ContentReduction in the Curriculum” press release, July16, 1998, www.moe.gov.sg/media/press/1998/980716.htm.

83. Teo Chee Hean (speech on national curricu-lum changes, October 8, 1999), http://www.moe.gov.sg/media/speeches/1999/sp081099.htm.

84. Colorado State University Department ofPublic Relations, “Colorado State University Wel-comes 40 Korean Teachers for One Month ofInnovative Science Education Training” (pressrelease, www.news.colostate.edu/Release/557, July15, 2005); and Indiana University, “Korean Teach-ers Taking in Differences, Making Notes DuringMonth-long Visit” (press release, newsinfo.iu.edu/news/page/normal/8580.html, July 30, 2008).

85. See, for instance, Masaru Tamamoto, “WillJapan Ever Grow Up?” Far Eastern Economic Review,(July 10, 2009), www.feer.com/essays/2009/july/will-japan-ever-grow-up; and Kathleen KennedyManzo, “Trends in Japan: Japan Continues Searchfor Academic Triumph,” Education Week, April 23,2008.

86. GDP per capita adjusted for purchasing-power-parity using International Monetary Fund, WorldEconomic Outlook Database, April 2009 edition,

www.imf.org/external/pubs/ft/weo/2009/01/weodata/index.aspx.

87. For instance, see Katherine Long, “WhichMath Book To Use? A Passionate Debate Rages,”Seattle Times, August 16, 2009, and Matthew E.Milliken, “Parents Blast ‘Reading Street,’” Herald-Sun, September 28, 2009.

88. Andrew J. Coulson, “On the Way to School:Why and How to Make a Market in Education”(book chapter prepared for The ClemsonUniversity Conference on School Choice ProgramDesign, forthcoming).

89. Diane Ravitch, The Language Police: HowPressure Groups Restrict What Students Learn (NewYork: Alfred A. Knopf, 2003).

90. Cornelia Dean, “Evolution Takes a Back Seatin U.S. Classes,” New York Times, February 1, 2005.

91. Scott Keeter and Juliana Horowitz, “OnDarwin’s 200th Birthday, Americans Still Dividedabout Evolution,” Pew Research Center, February5, 2009.

92. Wayne J. Urban and Jennings L. Wagoner Jr.American Education: A History, 3rd ed. (Boston:McGraw Hill, 2004), pp. 79–82.

26

RELEVANT STUDIES IN THE POLICY ANALYSIS SERIES

641. The Poverty of Preschool Promises: Saving Children and Money with theEarly Education Tax Credit by Adam B. Schaeffer (August 3, 2009)

629. Unbearable Burden? Living and Paying Student Loans as a First-Year Teacher by Neal McCluskey (December 15, 2008)

620. Markets vs. Monopolies in Education: A Global Review of the Evidenceby Andrew J. Coulson (September 10, 2008)

618. The Fiscal Impact of a Large-Scale Education Tax Credit Program by Andrew J. Coulson with a Technical Appendix by Anca M. Cotet (July 1, 2008)

616. Dismal Science: The Shortcomings of U.S. School Choice Research andHow to Address Them by John Merrifield (April 16, 2008)

605. The Public Education Tax Credit by Adam B. Schaeffer (December 5, 2007)

599. End It, Don’t Mend It: What to Do with No Child Left Behind by Neal McCluskey and Andrew J. Coulson (September 5, 2007)

STUDIES IN THE POLICY ANALYSIS SERIES

660. Lawless Policy: TARP as Congressional Failure by John Samples(February 4, 2010)

659. Globalization: Curse or Cure? Policies to Harness Global Economic Integration to Solve Our Economic Challenge by Jagadeesh Gokhale (February 1, 2010)

658. The Libertarian Vote in the Age of Obama by David Kirby and David Boaz (January 21, 2010)

657. The Massachusetts Health Plan: Much Pain, Little Gain by Aaron Yelowitz and Michael F. Cannon (January 20, 2010)

656. Obama’s Prescription for Low-Wage Workers High Implicit Taxes, Higher Premiums by Michael F. Cannon (January 13, 2010)

655. Three Decades of Politics and Failed Policies at HUD by Tad DeHaven(November 23, 2009)

654. Bending the Productivity Curve: Why America Leads the World in Medical Innovation by Glen Whitman and Raymond Raad (November 18, 2009)

653. The Myth of the Compact City: Why Compact Development Is Not the Wayto Reduce Carbon Dioxide Emissions by Randal O’Toole (November 18, 2009)

652. Attack of the Utility Monsters: The New Threats to Free Speech by Jason Kuznicki (November 16, 2009)

651. Fairness 2.0: Media Content Regulation in the 21st Century by Robert Corn-Revere (November 10, 2009)

650. Yes, Mr President: A Free Market Can Fix Health Care by Michael F. Cannon (October 21, 2009)

649. Somalia, Redux: A More Hands-Off Approach by David Axe (October 12, 2009)

648. Would a Stricter Fed Policy and Financial Regulation Have Averted the Financial Crisis? by Jagadeesh Gokhale and Peter Van Doren (October 8, 2009)

647. Why Sustainability Standards for Biofuel Production Make Little Economic Sense by Harry de Gorter and David R. Just (October 7, 2009)

646. How Urban Planners Caused the Housing Bubble by Randal O’Toole (October 1, 2009)

645. Vallejo Con Dios: Why Public Sector Unionism Is a Bad Deal for Taxpayers and Representative Government by Don Bellante, David Denholm, and Ivan Osorio (September 28, 2009)

644. Getting What You Paid For—Paying For What You Get: Proposals for theNext Transportation Reauthorization by Randal O’Toole (September 15, 2009)

643. Halfway to Where? Answering the Key Questions of Health Care Reformby Michael Tanner (September 9, 2009)

642. Fannie Med? Why a “Public Option” Is Hazardous to Your Health by Michael F. Cannon (July 27, 2009)