Public Disclosure Authorized
Public Disclosure Authorized
Public Disclosure Authorized
Policy,Reseaech,
and ExternalAffairs
iWORKING PAPERS
Educationand Employment
Populationand HumanResources
Department
The WorldBank
July 1991
WPS734
Curricular Content,
EducationalExpansion,
and EconomicGrowth
Public Disclosure Authorized
Aaron Benavot
Many academicians, politicians, and educators strongly believe
that knowledge, organized in school curricula and transmitted
through school systems, contributes to the economic strength of
nations. How valid is this claim?
The Policy, Rcsearch. and Extemal Affairs Complex distributes PRE Working Papers todisseminate the Findingsof wor in progrcss and
to encourage the exchange of ideas among Bank staff and aDlothers interested in development issues. These papers carry the names qf
the authors, reflect only their views, and should be used and cited accordingly. 1Thefindings, interpretations, and conclusions are the
authors' own. They should not be attuibuted to the World Bank, its Board of Directors, its managemcnt, or any of its member countries.
Plcy,Research,and ExternalAffairs
Educationand Employment
WPS 734
This paper-a product of the Education and Employment Division, Population and luman Resources
Department-is partofalargereffort in PRE toexamine the effects ofprimary education on development.
Copies are available free from the World Bank, 1818 H Street NW, Washington, DC 20433. Please coIntalct
Cynthia Cristobal, room S6-214, extension 33640 (27 pages, with tables).
Benavot examines whether national variations in
curricular content and subject area - as distinct
from growth in enrollment or qualitative provisions - have a significant impact on economic
development.
The study focuses on primary education in
60 nations and assesses the economic impact of
an emphasis on eight different primary level
subject areas, with special attention to matlhematics and science.
Benavot found that the curricular content of
mass education is directly related t", national
economic growth. This relations' :.p, however, is
not consistent across all subject areas and all
types of countries.
Countries requiring more hours of elementary science education, for example, generally
experienced more rapid increases in their standards of living during the periods from 1960 to
science education at the primary
1985. VWhcther
level is ,Ic key causal factor and wlhetlherthe
explicit content of the subject area is the key
mechanism remain unclear.
The design, reform, and study of national
school curricula are increasingly visible in
political and scholarly agendas. Conventionlal
wisdom on these matters, however, may cloud
rather than clarify a vision of tlhepotential
economic benefits of different choices of subjects for curricula.
The economic consequences of emphasizing
different subject areas should not be the sole
criterion for decisionmaking in designing
curricula. However, these consequences can
provide one useful element for promoting more
informneddiscussion among such interested
parties as parents, school administrators, national
and international planners, and educational
researchers.
The PRE Working Paper Series disseminates the findings of work undCr way in the Bank's Policy, Research. anid Extemal
Affairs Complex. An objective of the series is to get these findings out quickly, even if presentations are less than filly lx)ished.
findings, interpretations, and conclusions in these papers do not necessarily represent official Bank policy.
LThe
Produced by the PRE Dissemination Center
Curricular Content, Educational Expansion, and Economic Growth
by
Aaron Benavot
University of Georgia and HfebrewUniversity of Jerusalem
Table of Contents
Introduction
1
Curricular Content and Economic Development: Arguments and Evidence
2
Curricular Policy Reforms in Less-developed Nations
5
Data Sources
7
The Meaning of Official Curricular Data
9
Research Design and Methodology
11
Analyses and Findings
12
Discussion
16
References
19
Tables
23
This research was supported by grants from the National Science Foundation
(SES 85-12561). the World Bank, and by a Spencer Postdoctoral Fellowship from the
National Academy of Education. Earlier drafts of this paper were presented at the
Comparative and International Educational Society (March 1990) and at a seminar at
the Population and Human Resources Department of the World Bank (February 1990).
e
National-level studies of education and econom-icgrowth have established that mass
educational expansion -- mainly at the primary level but, to a lesser extent, at secondary level -- has
a significant positive impact on economic growth (Meyer et al., 1979; Wheeler, 1980; Walters,
1981; Benavot, 1989). There is also evidence that "qualitadive"features of national school systems
(e.g., pupil expenditures, textbook provisions, teacher training) have modest economic effects,
especially in the developing world (Heyneman and Siev-White, 1985; Fuller and Heyneman,
1989). Debate currently continues as to specific array of polidcal, social and historical conditions
under which the economic impact of mass education is most likely to be pronounced (e.g., Fuller et
al., 1986; Hage et al., 1988).
The present study, by contrast, focuses on a new line of inquiry. It explores, from a crossnational perspecdve, the relationship between official curricular policies and long-term economic
growth. Specifically, it seeks to determine whether, and to what degree, national variations in
curricular content and structure -- as distinct from enrollment growth or qualitative provisions -have a significant impact on economic development. Comparative educational research typically
analyses the content of national school systems according to a three-tiered classification scheme: the
official, intended curriculum; the implemented curriculum; and the achieved curriculum (Garden,
1987). This study exclusively focuses on the economic consequences of the first tier -- the formally
proscribed curriculum set forth by national educational authorities -- and only for primary
education. Based on a sample of over 60 nations (including 43 less-developed counties), it
examines the impact of official emphases in eight different primary-level subject areas (as measured
by average annual hours of instruction) on long-term economic growth between 1960 and 1985.
Mathematics and science education, two subject areas which many believe to be especiallypertinent
to the economic prosperity of nations, are highlighted throughout the paper.
The analyses and findings reported in this study are significant for severl reasons. First, they
provide a new perspective from which to evaluate the merits of national policies such as lengthening
the school term (e.g., Karweit, 1985) or requiring more instruction in academic as opposed to
vocational or "value-oriented" subjects (e.g., religion, moral education, civics). Second, they entail
a preliminary test of the widely assumed, though still unsubstantiated,claim that an emphasis on
l
mathematicsand scienceeducadonin elementaryor secondaryschoolsis a criticalprecursorof
economicprospaity. 1hird,they indirectlyaddressan issue of greatconcernto comparative
educationalresearch-- namely,what are the underliningmechanismsthroughwhichmass
schoolingplays such a pivotalrole in the socialandeconomicdevelopmentof Third Worldnations
(Colcough,1982;Psacharopolousand Woodhall,1985,Lockheedand Verspoor,1990)?
CurdrculaContentanidEconoic DevelopMent:
Arggmnts gnd vdence
Educationalknowledgeorganizedin the curriculumand transmittedby the public school
systemsignificantlycontributesto the economicstrengthand vitalityof a nation.Sucha view has
long been endorsedby manyacademics,politiciansand educators.Especiallyin the early
industrializingnations(England,Germany,the UnitedStatesandFrance)stateauthoritiesengaged
in protracteddebatesover the design,controland standardizationof schoolcurricula-- in part
becauseit was thoughtthat the contentof schoolingcouldplay a role in intemationaleconomic
competition(Goodson, 1983;Popkewitz,1987;Kleibard,1987;Hage et al., 1988;Cha, 1989).
Many believedthen that an economicallyrelevantschoolcurriculumemphasizingscientific
reasoning,mathematicalskills,modemlanguageproficiencyand technicalknowledge(ratherthan
high statusinstructionin classicallanguagesand literature)was neededin order to ensurea
productiveworkforceand a dynamicindustrialeconomy.
Similarbeliefsare voicedtoday. Recentcallsto re.orn nationalcurricularpoliciesare oftentied
to concernsover a nation'scompetitivenessin the worldeconomy-- or the lackthereof(Nationat
Risk, 1983).Considerthe followingtypicalargumentwith referenceto the U.S. and Japan which
is especiallysalientin the popularpressandin the speechesof NorthAmericanpoliticaland
businessleaders.Japaneseworkersaremoreproductive,manufacturehigherquality goods,and are
more committedto the economicenterprisethan AmericanworkersbecauseJapaneseschoolsare
more effectiveagenciesof socializationandtrainingthan Americanschools. Japanesestudents
learn morein schooland are betterdisciplinedin the classroomwhichspillsover into the
workplace; they are exposedto moreacademicsubjects,especiallymathematicsand science,which
facilitateshigherordercognitivedevelopmentamongpupilsas well as technologicalinnovationand
z
economic competitiveness. The decline of educational excellence in the United States, so the
argument goes, mirrors its declining share of international markets, its diminishing capacity for
manufacturing high quality goods and its rising trade deficits. If educational standards were raised
by teaching more "basics" (read: academic courses) and by assigning more hours of homework,
then achievement scores would rise, worker productivity would increase, scientific and
technological inventiveness would revive, and the United States would eventually reaffirn its
competitive edge in the world economy (among many such citations, see those discussed in Gross
and Gross, 1985; Doyle and Levine, 1985).
In addition, studies of cross-national variations in educational achievement have often been
used to partially explain past economic growth rates or foreshadow future ones. The findings from
two major comparative research projects are particularly relevant in this regard. The first involves
the well-publicized findings of recent lEA studies of academic achievementin the areas of
mathematics and science (McKnight et al., 1987; Rosier, 1987; IEA, 1988; Travers and Westbury,
1910); the second relates to the cross-cultural research of H. Stevenson and J. Stigler comparing the
academic achievement of Japanese, Chinese and U.S. elementary students (Stigler, Lee, Lucker and
Stevenson, 1982; Stevenson, Lee and Stigler, 1986). These two projects, taken together, tend to
show average or below average scores of U.S. and some Western European pupils but significantly
higher achievernents among Japanese and other Asian students on standardized mathematicsand
science tests.
Based on the (mostly) cross-sectional analyses performed in these studies, researchers have
concluded that national variations in academic achievementaxevery much related to the organization
and structure of the national school curriculum. In addition, the basic descriptive findings of these
studies have lent support to rhetorical arguments tying nationalcurricular content to economic
performance. Thus, without any solid empirical support, it has become received wisdom that the
composition and structure of national curricula, by their influence on student achievement levels,
have important effects on the quality and productivity of work force and, in the long run, on a
nation's competitive position in the world economy.
How valid and logically consistent are such claims? To begin, when the comparison is limited
to Japan and the United States, these arguments have some face validity. Japan has a much longer
3
school year than the U.S. (240 days versus an average of 180 days) arid, if "extra-school"
education and tutoring programs are added to the equation, the typical student in Japan is exposed to
significantly more formal instruction in "hard" academic subjects than the typical U.S. student
(Kobayashi, 1984). (This is not necessarily true for each subject area at each level of the educational
system). Greater levels of overall instructional ime, both in-school and extra-school, are certainly a
factor in the superior performance of Japanese pupils on mathematics and science tests and,
presumably, thev have contributed to the rapid expansion of the Japanese economy in recent
decades, among the fastest-growing in the world.
Notwithstanding its intuitive appeal, reasoning along these lines can be faulted on several
grounds. First, it ignores salient social and educatioaal differences between the two countries -- the
dominant pattern of social mobility, the degree of cultural homogeneity, the emphasis of effort over
innate ability and so on -- differences that may account for subsequent educational and economic
performance (Shimahara, 1985; U.S. Department of Education, 1987). Second, when a larger
sample of countries is examined, the relationship between total instructional time (or length of the
school year) ard national educational achievement weakens considerably, at least in the area of
mathematics (McKni ht et al., 1987).
Third, and perhaps most significantly, even if it canibe shown that actual instructional time, the
length of the school year, or other features of the national curriculum have a strong impact on levels
of student achievement, this does not necessarily corroborate the claim that student proficiency in
certain school subjects contributes to economic growth. The latter argument, which is based on
aggregate-level causal effects, cannot be conclusively validated by evidence of individual-level
effects (see Meyer, 1977). Furthermore, the fact that students, while in school, achieve high scores
in certain acadernic subjects does not necessarily mean that such demonstrable proficiencies persist
over the life-cycle or are effectively applied upon entering the workplace. Indeed, this presumption
of long-term socialization effects is especially suspect in countries where the credentials students
acquire carry more weight than the knowledge they may have actually gained through schooling
(Collins, 1979).
Conversely, if structural features of national school systems are shown to have weak or no
effects on academic achievement as measured by standardized tests, this does not necessarily mean
that the formerare unrelatedto nationaleconomicproductivity.For example,by increasing
students'interests in university-levelprogams in science,engineeringor technology,or by
influencingstmdentattitudestowardsand underlyingpublicsupportfor governmentinvestnent
polie&sin scienceand technology,on the one hand,or basicresearchand developmenton the
other,featuresof the schoolcurriculumcouldhave long-termdevelopmenteffectsin a contextof
relativelylow achievementlevels.In short,thereis a dearthof well-designedcomparativestudies
whichdirectlyaddressthe issue of whethervariationsin nationalcuricular policies,independentof
studentachievementlevels,have significant,aggregate-levelimpactson macroeconomicchange.
CurricularPolicy Reformsin Less-DevelopedNations
Whileargumentsconcerningthe economicimpactof schoolcurriculahave been intensely
debatedand discussedin advanced,industrializednations,they have specialrelevancein the context
of the developingworld.The schoolcurriculaof newlyindependentnations,whichwere largely
shapedby the interestsof formercolonialpowersand missionaryauthorities,have been re-assessed
-- and often restructured -- in light of a diverse array of new purposes: forging national unity,
traininghighlyskilledmanpower,spreadingliteracythroughoutthe populationand promoting
"modern"values andideas.Althoughmanydevelopingcountriesmadeonly minorchangesto the
previouslyused schoolcurriculaor simplyimportednewcurricularmaterialsfrom the metropolitan
centers,others undertooksignificantcurricularreformsto re-organizethe contentof mass
education.The push to integrate"modem"schoolsubjectswithscientificand mathematicalcontent
alongsideolder subjectemphasessuchas religion,languageandpracticaltrainingwas especially
pronounced.Duringthe post-WorldWar II era, for example,virtuallyall developingcountries
mandatedsome formof mathematicsand scienceinstructionin publicelementaryschools(Kamens
and Benavot,forthcoming).And at the lowersecondarylevel (grades7-9), thesetwo subjectareas
have becomeubiquitous,core elementsin the schoolcurriculaof developingcountries,takingup
closeto 40 percentof total instructionaltime (UNESCO,1986).
Table 1 presentsdetailedlongitudinalevidenceconcerningthe growingimportanceof
mathematicsand scienceeducationin the officialschoolcurriculaof less-developedcountries.In the
areaof mathematics,the top sectionsof table I showthat the relativeemphasison and the average
amountof instructionaltimedevotedto this subjectat the primarylevelincreasedin all developing
regions between 1960and 1980.Curriculartimeallocatedto scienceeducationalso increased,
althoughthe two developingregionswhichdevotedthe most instructionaltime to this subjectin
1960(LatinAmericaand Asia)slightlyreducedtimeallocationsin the 1980s.If anything,lessdevelopednationsplacedgm= curicular emphasison elementarysciencethanmore-developed
OECDnations.Interregionalvariationsat the primarylevelare apparent,but relativelysmall:
mathematicsand scienceeducationwere givensomewhatmoreemphasisin Latin Americaand the
Caribbeanand somewhatless emphasisin the MiddleEast and Norh Africa.Modestregional
variationin thesetwo subjectareasis alsonoticeablein the officialcuricula of lowersecondary
grades(see the bottomsectionof table 1).
The mair.pointof this table is that nationsaund the orld
have substanally ina
e
d
all in the ThirdWord.
instructionalime in mathemadcsand the sciences.the two
subjectareastought to have the ratest economicandtechnolscal zlevance, en thoughthere
is littleconclusivescientific-ev-idence
colfirmin2gthe lonL-termecnoi
imDat of thesesubjects.
Two ideas -- one, that it is necessaryto exposeyoungpeopleto the basicprinciplesof modern
science(the worldis empiricaland lawfuland govemedby naturalforces)and, two, that the
applicationof scientificknowledgethroughtechnologywill improvethe qualityof life of
individualsand contributeto community,nationaland economicdevelopment(UNESCO,1984)-have apparentlybecomeabidingarticlesof faithin the modemworld.
Thesecherishedbeliefs-- and the moregeneralthesisthatthe contentof schoolingdirectly
contributesto a rangeof societaloutcomessuchas economicgrowthor sciendficproductivity-have never, as previouslynoted,been empiricallytested. In the next sectionsinitialresulh,from
such a test are describedand reported.Usinga cross-nationaland longitudinalresearchdesign,the
analysesexplorewhethernationalvariationsin annualinstructionaltimedevotedto mathematicsand
scienceas well as six othersubjectareashad positiveeffectson economicgrowth.In additionthey
examinewhetherthe economicimpactof differentschoolsubjectsvariesamongmore-and lessdevelopedcountries.
D-ataSoumc
Mostof the cross-nadonaldata employedin this studywere takenfrom widelyavailable
nationalaccount statisticspublishedby the Ur.itedNations,UNESCO,and the World Bank. The
main dependentvariable,economicgrowth,was measuredby per capitareal grossdomestic
product(RGDP)in constant 1980internationalprices (Kravis,Hestonand Summers,1982;
Summersand Heston, 1988).1This variable,in contrastto the more widely-usedmeasuregross
nationalproduct(GNP),providesmorerealisticandcomparableestimatesof nationalincomelevels
and purchasingpower parities.Sincethe distributionof per capitaRGDPdata is positivelyskewed,
like most other measuresof economicdevelopment,it was loggedin all the analysesreported
below.
Apart from the mainexplanatoryvariableswhichrefer to annualinstructionalhoursin different
curricularsubjects(to be discussedbelow),severalotherindependentandcontrolmeasureswere
incorporatedinto the analyses.Two measuresof the expansionof masseducationwere employed-the primaryand secondaryenrollmentrates,respectively-- whichrefer to the proportionof the
relevantschool-agepopulationenrolledin the lower two levelsof the educationalsystem
(UNESCO,1970).In addition,three othercontro!measureswere includedsincepreviouscrossnationalstudieshave shownthem to have significantlyeffectson economicgrowthrates.These
include:1) populationgrowth: this was measuredby each nation'stotalfertilityrate (UN, 1986)
and was expectedto have negativeeffectson economicdevelopment;2) exportcommodity
concentration:this indicatormeasuredone aspectof economicdependenceandwas expectedto
negativelyaffect economicgrowth(WorldBank, 1983);and 3) a "dummy"variablefor nations
whose economiesare highlydependenton oil and mineralexporting,whichwas thoughtto have a
positiveeffect on economicgrowth.
Data on nationalcurricularstructureandcontentcame froma wide arrayof publishedand
unpublishedsources.A considerablearnountof curricularinformationwas gatheredfromreportsof
internationalagenciessuch as UNESCO,comparativeeducationalhandbooksand encyclopedia,
historicalsurveysof nationaland colonialeducationalsystems,contemporarycase studiesof
educationaldevelopmentsin particularcountries,and two sets of microfichefrom the Intemational
1. Similiarresults wereobtained,thoughfor a slightlysmallersample,whenper capita
GNP was employedas the dependentvariable.
7
Bureauof Education(IHE,1984;1986).Data on schoolcuricula was alsocoded fromofficial
correspondenceand naional publicationsthat were receiveddirectlyby a team of researchersof
whichthe author was a member.Of the over 150nationalministriesof educationto whichshort
questionaireswere sent in 1985,approximatelyhalf sent replies,of varyinglengthand detail,with
infofmationon curricularpoliciesand plans. A detailedlistingof all the sourcesusedin this study
can be found elsewhere(Benavotand Kamens,1989;Kamnens
and Benavot,forthconiing).
Two basicpiecesof inf3rmationwerecoded from thesesources: first, a list of the subjects
taughtduringthe elementaryschoolcyclein eachcountry;and second,the numberof school
periodsor school"hours"devotedto each subjectduringthe primaly schoolcycle.2 Thiscoding
procedureproducedtwo types of variablesfor eachsubject: first,a dichotomousvariablewhich
simplynoted whetheror not a particularsubjectwas taughtin the officialcurriculum;and second,a
continuousintervalvariableindicatingthe proportionof totalschoolperiodsor "schoolhours"that
each subjectwas taughtduringthe first six yearsof primaryeducation.Data on thesevariableswere
then groupedaccordingto historicalperiod(e.g., 1945-1969;1970-1986)dependingon whenthe
codedcurriculartimetablewas in effect.
In additionto thesetwo curricularmeasures,a third,more"refined"indicatorof official
curricularemphaseswasdeveloped.Insteadof measuringthe .-elativeemphasisof each subjectin
the nationalcurriculum(i.e.,the proportionof totalcurricularperiodsdevotedto mathematicsor
scienceeducation),it estimatedthe actualamountof time (in annualhours)that pupilswereexposed
to instructionin eachsubjectarea duringeachyear of the primaryschoolcycle.Recentreports
(e.g., McKnightet al., 1987:53; UNESCO,1986)as well as past ones (e.g., Dottrens, 1962)
2. Eachcurricularsubjectlistedin an officialtimetablewas codedinto oneof eight general
categories: CombinedLanguageeducation(readingand writingin national,official,local,
foreignand/orclassicallanguages),Mathematics,Sciences,CombinedSocialSciences
(socialstudies,history,geography,civics),CombinedMoraland ReligiousEducation,
AestheticEducation(music,arts, drawing,dance),PhysicalEducation,and Pre-Vocational
Educationor P tical Subjects(manualtraining,industrialarts,agriculture,domestic
science,vocationaleducationand business).Curriculartime in electivesubjectsandin
various "peripheral"activities(e.g.,hygiene,recreation,extracurricularactivities,recess,
miscellaneous)werecoded as "other"and not incorporatedin the present study.A
combinedcategorymeansthat instructionwasofferedin eitherone or all of the specific
curriculartopicslistedin that category.
suggestthat cross-nationalvariationin annualinstructionaltime is quite marked.Thus, for example,
whilethe relativeemphasison mathematiccor sciencein the officialprimarycurriculummay change
slowlyover time or vary quite modestlyover nationalspace,measuresof the numberof hours
mathematicsand sciencesubjectsare taughteach year are thoughtto varymore markedly. Also,
manyscholars.naintainthat studentslearnmorewhenthey are exposedto moreinstructionin a
particularsubjectarea or whenthey spendmore time "on task"(Fisherand Berliner,1985;Karweit,
1985;Murphyand Hallinger,1989). Onemight then draw thefollowingtwo inferences:1) nations
gain more, in economicor politicalterms,whentheirmass schoolsystemsrequiremoreoverall
time of fornal instrucdion;and 2) nationsgain morewhen they exposegreaternumbersof students
instruction.The firstinferenceassumesan additiveeffect of
to greateramountsof formnal
instructionaltimewhereasthe secondone assumesa multiplicativeeffect.Both ideaswilUbe tested
in this study.
To developestimatesof annualinstrucdonaltime,publishedsourcesand official
on the lengthof the schoolyear (in
correspondencewereexaminedso as to retrieveinformaadon
weeksor days), the lengthof the schoolweek (in weekly schoolperiodsor actualhours of
instruction),and the lengthof a schoolperiod (in fractionsof hours).Combiningpreviouslycoded
curricularinformationwiththis newinformation,estimatesof averageannualhoursof instruction
fo: each subjectareawere constructedfor about80 nationsin the 1945-69periodand some 100
nations in the 1970-86period.These figuresrefer to the numberof forrnalinstructionalhoursthat a
typicalpfimaryschoolstudentis supposedto receiveeachyear in a givensubjectarea.Whenever
possible,schooltime set aside for recesses,breaks,assembliesand the like was excluded.
Althoughsome erroris inherentin this sort of data, thereis goodreasonto believethat they permit
reasonablyvalid andreliablecomparisonsacrossnationalspace(for furtherdiscussionof how these
figures were estimated,see Kamensand Benavot,forthcoming).
Th Meaningof OfficialCuricular Data
The curriculardataemployedin thisstudyare officialpolicydeclarationsof subjectsto be
taught in primaryschools,generallyproducedby governmenteducationofficialsin the form of
q
nationaltimetablesand syllabi.It is clearthat officialstatementsof curricularpolicycan be -- and
manytimes are - poorlyrelated to localeducationalpractices.This "slippage"or "loosecoupling"
betweenthe intendedand the implementedcurriculumhas beennotedin bothcentralizedand
decentralizededucationalsystemsin theindustrializedworld (Weick,1976;Meyerand Rowan,
191"; Baker and Stevenson,1989)and is assumedto be even greaterin the newlyestablished
educationalsystemsof developingnations(Hawes,1979;Lockheedand Verspoor,1990).
If the gap betweeneducationalrulesand classroomrealityis so pronounced,why study the
effectsof officialcurricularrules?First,an(' most obviously,officialstatementsof whatchildren
widelyunderstood,both intemallyand externally,as
are to studyreflectnationalcommnitments
ofricialdefinitionsimpingeon the fonnalorganization
carying authoritativeintent.At a mninimum,
of local sclhoolsystems.Certainbodiesof schoolknowledgeare officiallyauthorized;othersare
not. The approvedcontentareasare distributedthroughoutthe daysand yearsof the schoolingcycle
accordingto relativelyexplicitgoals.Nationaldifferencesin officialtimeallotmentsto particular
subjectareasreflectgenuinedifferencesin subjectemphasesandeducationalphilosophies.Second,
under certainconditions,officialcurricularpoliciesmay closelyapproximatethe actualcontentof
local schools.For example,when thereis strongconsensusbetweenlocal teachersandnational
administratorsover what shouldbe taughtin the classroomor whenlocal schoolsencountertight
administrativecontrolsover classroompracticesand procedures(see Bakerand Stevenson,1989),
the gap betweenthe officialandimplementedcuriculum may be relativelysmall.
But do officialdataobscuremoie than theyreveal?If officialpoliciesconcerningthe
curriculumare simplyideologicalstatementsintentionallydevelopedby nationaleducationalleaders
to reflectthe "proper"educationalstandardsof intemationaldonoragencies(whomakecritical
decisionsregardingeducationalloansand grants),then our datamay show no morethan superficial
patternsof worldwideconformity.To be sure,nationaltimetablesand curricularstandardsare
deternined, in part, by extemal worldwideconventions,thoughnot necessarilythose embeddedin
internationaldonor agencies.Recenthistoricalresearchsuggeststhat therewas considerable
similarity(convergence)in officialcurricularpoliciespriorto the establishmentof international
educationalanddevelopmentagenciessuchas UNESCO,IBE and the WorldBank (Benavotet al.,
1988;Kamensand Benavot, 1990).
10
However, official curricular policies among contemporary nation-states are more than
superficial reflections of curricular standards imposed by extemal agencies. There is growing
evidence that the curricular policies articulated by national educational leaders are both reflective of
and influenced by an emerging global educational culture (see Meyer, Benavot and Kamens,
forthcoming). At the very least, it is interesting to examine whether national standards of curricular
policy have long-tern effects on important societal changes such as economic growth, and we shall
turn to this issue presently.
Research Design and Methodoogy
This quantitative cross-national study attempts to establish causal relafions between variable
features of social units (in this case, nations) using a non-experimentalresearch design. The
defining feature of this type of comparative research is its use of attributes of macrosocial units in
the explanatory statements (Ragin, 1987). As in previous cross-national analyses of economic
growth, this study uses aggregate, national level data. The unit of analysis is countries; the sample
size depends on the data coverage of the variables included in the models.
This study employs a multiple regression methodology using panel data, the advantages of
which have been discussed by Hannan (1979) and Tuma and Hannan (1984). Panel regression
analyses incorporate both variation between nation-states and variation over time. They basically
involve regressing the dependent variable at a latter time point (2) on the same variable measured at
an earlier time point (tl) and on a series of independentvariables also measured at the earlier time
point (tI). In the present study, a 25-year panel design was used in which the level of economic
development at 1985 was regressed on itself at 1960 (the "lagged dependent variable"), on
measures of instructional time devoted to different school subjects at 1960, and on several control
variables also measured about 1960.3
Analyses were performed separatelyfor all nations and for a smaller sample of only developing
nations in the Third World to check for differential effects. Since data coverage on most of the
above measures (except for curricular emphasis) is quite extensive, the "sample" of countries from
3. A shorter panel design in which fifteen year lagged effects were examined produced
results that are nearly identical to those reported in the paper.
11
differentworldregions,exceptsub-SaharanAfrica,is fairlyrepresentative.AppendixA lists the
specificcountriesincludedin the analyses.
Analysesand Findings
This sectionpresentsresultsfrom cross-nationalanalysesof curricularsubjectareasand their
effectson econornicgrowth.It beginsby brieflysurveyingdescriptivestatisticsconcemingthe
main explanatoryvariablesusedin the panelregressionmodels.Table 2 showsthat in 1960nations
required,on the average,870 hours of formalinstructionduringeach year of primaryschooling.
Overallhoursof instructionwere higheramongdevelopedcountries(917.0 hoursper year)than
the three R's tendedto takeup
among developingcountries(849.3 hoursper year). Furthermnore,
abouthalf of all instructionalhours: readingand writing(i.e., languageeducation)was allocated
over 300 hoursof instruction;arithmeticand othermath-relatedsubjectsabout 145hoursof
instruction.Theremaininginstructionaltimewas distributedamonga seriesof "core"subjects
whichwererequiredin almostall nationalprimarycurriculaanda set of more "peripheral"subjects
whichwere requiredin less than three-fourthsof all countries(seeBenavotet al., 19w). Core
subjectsincludedart and music (81.2 hoursper year), socialsciences(75.6hrs./year),natural
sciences(63.3 hrs./year),and physicaleducation(58.7 hrs./year).Peripheralsubjectsincluded
religionand moral education(50.5 hrs./year)and pre-vocationaleducation(65.4hrs./year).
Differencesin averagetime allocationsby levelof developmentwere minimalin scienceeducation
and religiousand moraleducation,as well as in the combinedareaof history,geographyand social
studies.In the remainingsubjectareas,developednationsplacedgreateremphasison aesthetic
educationandphysicaleducationandless emphasison pre-vocationaleducationthandeveloping
nationsin the ThirdWorld.
Table 3 presentszero-ordercorrelationsamongall the variablesusedin the panelregression
analyses.By and large,correlationsbetweenmost curricularemphasesand levelsof economic
developmentare weak and non-significant.Twosubjectareas,mathematicsand physicaleducation,
have positive,but not especiallystrong,associationswith economicdevelopment;pre-vocational
educationand economicdevelopmentare negativelycorrelated.As expected,associationsbetween
the otherindependentvariables(ratesof educationalexpansion,fertilityrates,exportcommodity
12
concentration)and economicdevelopmentare strongand in theexpecteddirection.Althoughtable 3
reportsseveralinstancesof relativelyhighcorrelationsamongindependentvariables,whichcould
potendallybe a sourceof multicollinearity(Hanushek,1979),thesewere later checkedusinga
series of regressiondiagnostics(Belsey,Kuh and Welsch, 1980)and found to have no adverse
impact on the mainpatternof resultsreportedbelow.
In the panel regressionanalyses,threedifferentmodels wereestimated.In the first model,a
baselineequationwas estimatedin whichlevelof economicdevelopmentin 1985wasregressedon
levelof economicdevelopmentin 1960,the primaryand secondaryenrollmentratesfor 1960and
the aforementionedcontrolvariables.The secondmodel addedto this baselineequationa single
measurefor overallyearlyinstructionaltimeat theprimarylevel(thatis, acrossall subjectareas).
The third modelestimatedthe economiceffectsof the moredetailedmeasuresof annualinstructional
timein eight differentsubjectareas(mathematics,science,language,art/music,social science,
religion/moraleducation,physicaleducationandpre-vocationaleducation)aftercontrollingfor the
influenceof educationalexpansion,ferility rates,high incomeoil exportersandcommodity
concentration.Thesethreemodels werefirst estimatedfor the full sampleof countries(table4) and
then for a smallersampleof less-developedcountries(table5). In all modelsonly unstandardized
regressioncoefficientsare reported;instancesin whichtheseregressioncoefficientsare either 1.5
timesor twice theirstandarderror are reportedas significant.
The economicimpactof interactioneffectsbetweencurriculartimeallotmentsandthe levelof
educationalexpansionat the primaxylevel werealso analysed(seetables6 and 7). Suchinteraction
(or multiplicative)effectsmeasurewhat percentageof studentsin each nationwere exposedto how
many hoursof fonnal instructionin each subjectareaand wereconstructedby multiplyingthe
primaryenrollmentrate by the annualhoursof instructionin eachsubjectarea.Theseindicatorsare
analogousto measuresof the "yield"or "productivity"of an educationalsystemwhichhave been
employedrecentlyin the IEAstudiesof mathematicsachievement(McKnightet al., 1987:61).
The baselinemodelestimatedin table4 (equationA) showsthat educationalexpansionat both
the primaryand secondarylevelshas a significantpositiveimpacton economicgrowthbetween
the resultsof earliercross-nationalstudies(Meyeret al,
1960and 1985.This findingconfwrms
1979;Benavot, 1989).Eachof the controlvariablesis significantand in the expecteddirection
13
except for the total fertility rate: not only is this indicator of population growth insignificant but, in
contrastto previous cross-national studies of economic growth, the coefficient is positive. This may
be due to the pardcular sample of countries included or, more likely, to the historical period under
examination.
Table 4 also reports that the effect of overall annual hours of instruction across all subject areas
is positive but not significant (equadon B). When this measure is broken down by subject area
(equation C), two of the eight subject areas have significant effects on economic growth: science
education (a positive effect) and pre-vocational or pracdcal education (a negative effect).
Interestingly, there is no evidence that instructionalhours devoted to elementary mathematics have a
significant impact on economic growth -- indeed, the direcdon of the regression coefficient
associated with mathematics education is negative.
Does this pattern of curricular effects change when a smaller sample of less-developed nations
is examined? For the most part, no. Table 5 shows that overall hours of instruction continues to be
positive but non-significant,that the effects of science education and pre-vocational education are in
the same direction and still significant, and that mathematics education continues to have little or no
impact on economic growth. The main difference for less-developed countries is, surprisingly, with
respect to aesthetic education and physical education. The results reported in equation C of table 5
indicate that Third World countries which taught more hours of art and music education
experienced, other things being equal, stronger growth rates; those that devoted more time to
physical education experienced weaker growth rates.
The reasoning behind these unexpected relationships is not entirely clear. Greater emphasis on
physical education in some countries may reflect a formnof paramilitary training and a welldeveloped military infrastructure which, in the long run, may drain potentially productive resources
from the overall economy. Greater emphasis on art and music education may indicate the cultural
impact of Western educational conceptions and philosophies which have historically placed a
prenium on this form of child socializationat the elementary level. If such systems have been
penetrated culturally by the West, then their economies may also be oriented to Western market
arrangements and thus more likely to have been targets of internationaleconomic aid and loans.
1.-
These patterns,while suggestive,certainlydemandfurther analysis.
The final seriesof panelmodelsexaminethe muldplicative(or non-additive)effectsof
enrollmentexpansionand hoursof instructionin differentsubjectareason economicgrowth(see
tables 6 and 7). The basicfindingsfrom thesemodelstend to reinforceand, to a lesserextent,
expandon those reportedfrom the additivemodels.First, the resultsindicatethat exposinga larger
primaryschoolcohortto morehoursof scienceeducationproducedclear benefitsin termsof
economicgrowthbetween 1960and 1985.Second,nationswith greaterpropordonsof enrolled
primaryschoolstudentswho wereexposedto morehoursof practicalor vocationalsubjectmatter
experienced,as notedearlier,a significantnegativeeconomicimpact.Third,therewas no evidence
of any interactioneffect of mathematicseducation.This threeresultsobtainedin boththe full sample
of 63 countriesand in the smallersampleof 43 less-developedcountries.Furthermore,the positive
impactof instructionalhors devotedto artsand musicin less-developednationswas also apparent
in thesemultiplicativemodels (seeequationC of table7).
Two additionalinteractioneffectswere significantand deserveto be noted.First,in table 6, the
modelestimatedin equationB indicatesthat amongnationswith highprimaryenrollmentratres,
overallinstructionaltimehad a positiveand significantimpacton economicgrowth.Irrespectiveof
the subjectstaught,moreclass time in expandedmass schoolsystemsproducedlong-term
economiceffects.Second,in equationC of the same table,there was, for the first time, evidence
that annualhours of languageinstructionin moreexp2ndedprimaryschoolsystemshad a positive
economicimpact.This interactioneffect appearslimited,however,to FirstWorld nations: while
the effect of languageeducationis positiveandsignificantin table6, it is positive,but nonsignificantin table 7 whichexclusivelylookedat less-developednations.
Overall,theseresultsindicatethat mostsubjectsareasof the primaryschoolcurriculum-- at
leastwhen measuredin terns of officialannualhoursof instruction-- are unrelatedto long-tenn
changesin nationalecononic production.Twosubjectsareas,however,were foundto have
consistentand significanteffectson econonmcgrowthduringthe periodunderexamination:annual
hoursof instructiondevotedto scienceeducationhad strong,positiveeffectsin both additiveand
multiplicativemodelsas well as in both of the samplesexamined.By contrast,annualhoursof
instructiondevotedto pre-vocationalsubjectmatter(farming,manualtraining,agricultue,domestic
science, etc.) had strong, negative effects in each of the specified models.4 Possible explanations
for this pattem of effects will be discussed in the final section of the paper.
5slL5
Most educational researchers, if asked to evaluate the relative economic value of different
curricular subject areas, would undoubtedly rank mathematics and science education near the top,
assuming that the skills, knowledge and underlying principles imparted by these curricular subjects
have the greatest relevancy to a nation's economy and productive potential. Many would also
underscore the basic importance of reading, writing and communication skills (for future workers in
the labor market) which presumably are taught and stressed in the area of language education. The
economic value of pre-vocationalor practical subject matter (from manual training to fanning and
domestic science) would probably be seen in a more controversial light: some would argue that
such areas of the curriculum have direct spillover effects for local and national economies,
especially in the Third World where primary education in the final destimation of most students;
others would question their relevance since much of what is taught may become quickly outmoded
due to technological changes or since this subject area is viewed by many students, parents and
teachers as marginally important to the overall educational enterprise and to subsequent occupational
attainment. The remaining curricular subjects (i.e., art, music, history, geography, social studies,
religion, moral education, and physical education), while possibly important for the cognitive and
moral development of young children or for a country's sense of national identity, cultural
cohesiveness, common political purpose and overall national integration, would not be viewed as
directly relevant to national economic prosperity.
The positive impact of science education and the negative impact of pre-vocational apparently
provide credible evidence for the promoters of science, especially those who would increase its
emphasis in elementary schools, and the detractions of vocational education who would reduce
4. Additional analyses, not repoited here, indicate that the positive impact of science
education is strong throughout the Third World, especially so in some of the poorest and
least developed nations of the world. On the other hand, it was found that while prevocational education had negative effects in both poorer and richer LDCs, they were only
significant in the first category of nation-states.
1(b
substantially its place in the primary curriculum. Such an inference, however, would be a narrow
and overly simplistic reading of the findings. Several altemative explanations can be advanced (and.
indeed, need to be tested).
For example, it is quite possible that the measure, annual instructional hours in elementary
science, is actually a proxy variable for national emphasis on science education in secondary
schools or university institutions or for basic national commitments to scientific research and
development. Whether the "real" economic impact of science education is to be located at another
level of national educational systems, rather than at the elementary level, needs to be explored
systematically by employing controls for national emphases on science education and scientific
research and development.
Another potential explanation, especially in relation to the developing world, concems what the
content of science education may actually signify and what transformative role this curricular subject
may entail for the pupils who are exposed to it. Of all the curricular subject areas in the primary
curriculum, science education is most closely associated with a Western, rationalistic, cause-andeffect view of the world. For students growing up in agrarian economies and strong religious
traditions and with minimal contact to modern medical practices and new technologicaldevices,
science education at the elementary level may be their first exposure to a fundamentally different
view of the world. If such be the case, then the economic impact of science education may have
more to do with the "hidden" cultural rules and orientations being transmitted, rather than the
specific fornal scientific content being taught. Such an interpretation, while plausible, would be
difficult to establish within an exclusively quantitativeresearch design.
In sum, this paper has established that the curricular contenitof mass schooling, as distinct
frim its quantitative expansion, is related to national economic growth, although not for all subject
areas usually thought to be economically relevant and not across all types of countries. Countries
requiring more hours of elementary science education, other things being equal, experienced more
rapid increases in their standards of living during the 1960-1985period. Whether the emphasis on
science education at the primary level is the key causal factor and whether the explicit (rather than
implicit) content of this subject area is the key mechanism at work was not established and demands
further exploration.
17
Future research should also cast a potentially critical eye on the fact that mathematics and
language education, at least at the primary level, appears little related to long-term economic growth.
Why this should occur, given the numerous vocal and passionate arguments advanced throughout
the world for the expansion of reading, writing and computation skills (i.e., literacy and numeracy)
among the mass population as an important precondition for internationalcompetitiveness and
economic development, is a surprising result emerging from this study. Equally interesting, in
terms of its political and educational ramifications, is the potential importance of aesthetic education
(i.e., art, music, dance, drawing) for the economic expansion of certain less-developed nations.
While the design, reform, and study of national school curricula are increasingly visible in both
political and scholarly agendas, received wisdom and common sense ideas in this area mnaycloud,
rather than clarify, our vision as to the potential economic payoffs of different curricular subject
areas. The relative economic consequences of different subject areas are certainly not the only
standard for deciding whether to emphasize or de-emphasize particular areas of curricular
knowledge, but they can provide a backdrop for a more inforrned discussion among all interested
parties: parents, local adminstrators, national and international planners, and educational
researchers.
Baker, David and David Stevenson (1989) "Consequences of Macro Educational Characterstics
for Micro Educational Processes: A Comparative Analysis of State Control of Currculum",
paper presented at the World Congress of Comparative Educadon, Montreal, Canada, June.
Belsley, David A., Edwin Kuh and Roy E. Welsch (1980) Regession Diagoostcs: 1denfifyiog
Influential Data and Sources of Collinearty. New York: Wiley-Interscience.
Benavot, Aaron (1989) "Education, Gender and Economic Development: A Cross-national Study."
Sociology of Education 62: 14-32.
Benavot, Aaron, David Kamens, John Meyer, Suk-Ying Wong and Yun-Kyung Cha (1988)
"World Culture and the Curricular Content of National Education Systems, 1920-1985."
Paper presented at the American Sociological Association annual meeting, Atlanta, GA.
Benavot, Aaron, Yun-Kyung Cha, David Kamens, John Meyer and Suk-Ying Wong (1990)
"Knowledge for the Mass Population: World Models and National Curricula, 1920-1986."
Manuscript under review. Department of Sociology, University of Georgia.
Benavot, Aaron and David Kamens (1989) The CurnicularContent of PrimanyEducation in
Developing Countries. PPR Working Paper 237, World Bank, Washington, D.C.
Cha, Yun Kyung (1989) "The Origins and Expansion of Primary School Curricula in the West:
1800-1920." Unpublished manuscript. School of Education. Stanford University.
Cha, Yun-Kyung (forthcoming) "The Effect of the Global System on Language Instruction in the
Curriculum, 1850-1986." Sociology of Education.
Colcough, Christopher (1982) "The Impact of Primary Schooling on Economic Development: A
review of the evidence." World Development 10: 166-185.
Collins, Randall (1979) The Credential Society. New York: Academic Books.
Dottrens, Robert (1962) The Primary School CurriculLm. Paris: UNESCO.
Doyle, Denis and Marsha Levine (1985) "Business and the Public Schools." Phi Delta Kappan
(October): 113-118.
Fisher, Charles and David Berliner (1985) Perspectives on InstructionalTime. New York:
Longman.
Fuller, Bruce, J. Edwards and K. Gorman (1986) "When Does Education Boost Economic
Growth." Sociology of Education 59: 167-81.
Fuller, Bruce and Steven Heyneman (1989) "Third World School Quality." Educaional
Reseache. 18: 12-19.
Garden, R. A. (1987) "The Second IEA Mathematics Study." Comparaive Education Revie 31:
47-68.
Goodson, Ivor (1983) School Subiects and CLrriculumChan.
London: Croom Helm.
19
Gross, Beatriceand RonaldGross (eds.)(1985)The Great ScboolDebate.New York:Touchstone
Books.
Hage,Jerald, M. Gamier and B. Fuller (1988)"TheActiveState, Investmentin HumanCapital
and EconomicGrowth."AmericanSocological Reyiew53: 824-37.
Hannan,Michael(1979)"Issuesin PanelAnalysisof NationalDevelopment."Pp. 17-36in J.
Meyer and M. Hannan(eds.) Natinaljdeontad
the World System.Chicago:
Universityof Chicago Press.
Hanushek,Eric A. and John E. Jackson(1977)
stical Medj
i!thsoSocialSientsts.
New
York: AcademicPress.
Hawes,Hugh et al. (1979)Currculum and RealItyin Afcan Pmaty Schools.Bristol,UK:
Longman.
Heyneman,Stephenand D. Siev-White(1985)The L_ality
of Educationand Economic
Iye1pment. New York: Oxford UniversityPress.
IEA(InternationalAssociationfor the Evaluationof EducationAchievement)(1988)Science
Achievementin SeventeenCountries.Oxford:PergamonPress.
InternationalBureauof Educationand UNESCO(1984)Natonal Reportsto Intem2tonal
Conferenceon Education.39th Session(microfiche)Geneva:IBE and UNESCO.
IntemationalBureauof Educationand UNESCO(1986)Natonal Repos to nt-eratgnal
Conferenceon Education.40th Session(microEfiche)
Geneva:IBE and UNESCO.
Kamens,Davidand Aaron Benavot(forthcoming)"EliteKnowledgefor the Masses:The Origins
and Spreadof Mathematicsand ScienceEducationin NationalCurricula."AmericanJgral
oLfEducadon.
Karweit,Nancy(1985)"ShouldWe Lengthenthe SchoolTerm."Educatonal Researches 14:9-15.
Kliebard,Herbert(1986)The Strugglefor the AmericanCrriiculum. 1893-1958.Boston:
Routledgeand KeganPaul.
Kobayashi,V. (1984) "Japaneseand U.S. CurriculaCompared."Pp. 61-95 in W. Cummings,E.
Beauchamp,S. Ichikawa,V. Kobayashiand M. Ushiogi(eds.) EducationalPoliciesin
Crisi. New York:Praeger.
Kravis, Irving,Alan Hestonand RobertSummers(1982)World Productand Income:Interational
Comparisonsof RealGross Product.Baltimore:John HopkinsUniversity.
Lockheed,Marlaineand AdriaanVerspoor(1990)ImprovingPrimaryEducaion in Developing
Countries.Washington,D.C.:The WorldBank.
McKnight,Curtis,Joe Crosswhite,John Dossey, EdwardKifer,Jane Swafford,Kenneth
Travers,ThomasCooney (1987)The UnderachievingCurriculum:AssessingU.S. School
Mathematicsfrom an InternationalPerspective.Charnpaign,Ill.: StipesPublishingCo.
20
Meyer,John (1977)"The Effectsof Educationas an Instituton."AmericanJoural of Sociglo2y
83: 55-77.
Meyer,John and BrianRowan(1977)"Institutionalized
Organizations:Formal Structureas Myth
and Ceremony."ArnericanJournalof Sociology83:440-463.
Meyer,John, MichaelHannan,RichardRubinsonand GeorgeThomas(1979)"National
EconomicDevelopment,1950-70,"In J. Meyer and M. Hannan(eds.) Natiga
Developmentand the World System.Chicago:Universityof ChicagoPress.
Morawetz,David(1977)Twen1y-FiveYearsof EconomicDevelopment.Baltimore:John Hopkins
UniversityPress.
Murphy,.Tosephand PhilipHallinger(1989)"Equityas Accessto Learning:Curricularand
instructionaltreatmentdifferences."Journalof Curricululm
Stdies 21: 129-49.
A Nationat Risk (1983)NationalCommissionon Excellencein Education.
Popkewitz,Thomas(ed.) (1987)The Formationof the SchoolSubjects.New York:FalmerPress.
Psacharopoulos,Georgeand MaureenWoodhall(1985)Eduicationfor DeQQlnment.New York:
Oxford UniversityPress.
Ragin,Charles(1987)The ComparativeMethod.Berkeley:Universityof CaliforniaPress.
Rosier, Malcolm(1987)"The ScienceInternationalScienceStudy."
mQdnpatie
Education
Review31: 106-128.
Shimahara,Nobuo (1985)"JapaneseEducationand Its Implicationsfor U.S.Education."fbi
Delta Kapan 67: 418-21.
Stevenson,Harold, S. Lee and J. Stigler(1986)"MathematicsAchievementof Chinese,Japanese
and AmericanChildren."Sgienc 231:693-99.
Stigler, James,S. Lee, G. Lucker,H. Stevenson(1982)"Curriculumand Achievementin
Mathematics."Journalof EducationalPsychology74: 315-22.
Summers,Robert and Alan Heston(1988)"A New Set of InternationalComparisonsof Real
Productand Price LevelsEstimatesfor 130 Countries:1950-1985."The Reviewof Incom
and _Wealth
34: 1-25.
Travers,Kennethand Ian Westbury(1990)The IEAStudy of MathematicsI: Analysisof
MathematicsCurricula.Oxford:PergamonPress.
Tuma, Nancy and MichaelHannan(1984)SocialDynamics:Modelsand Methods.Orlando,FL:
AcademicPress.
United Nations (1986)World PopulationTrends,EstiMatesand Projectionsas Assessedii 1984.
New York: Departmentof InternationalEconomicand SocialAffairs,UN.
UNESCO(1970)StatisticalYearbook.Paris: UNESCO.
UNESCO(1986)The Placeof Scienceand Technologyin SchoolCurricula:A GlobalSuy.
21
Paris: UNESCO.
UNESCORegionalOfficefor Educationin Asia and the Pacific(1984)SciefnceEducan inAsia
nd Eadfi . Bangkok:UNESCO
U.S. Dept. of Education(1987)IJUS,Study of Educationin Japan.Washington:GPO.
Walters,Pamela Barnhouse(1981)"EducationalChangeand NationalEconomicDevelopment."
Hvard Educaional Review51: 94-106.
Weick,Karl (1976)"EducationalOrganizationsas LooselyCoupledSystems."Adminisave
Science,Ouarterly21: 1-19.
Wheeler,David(1980)HumanResourceDevelopmentandEconomicGrowtlhin LQDC's.
World
Bank Staff WorkingPaper No. 407. Washington,D.C.: World Bank.
Wong, Suk-Ying(forthcoming)"The Evolutionand Organizationof the Social Science
Curriculum:A Cross-NationalStudy: 1900-1986."Sociologyof Education.
World Bank (1983)World Tables.Third Edition.Washington:WorldBank.
22
Table 1:
Curricular Emphaseson Mathematicsand ScienceEducationat the Primary
and Lower SecondaryLevels,1960s° 1980s,by World Region
(constantcases)I
A. Proportionof the OfficialPfimaryCurticulun Devotedto Mathematicsand ScienceEducation
LatinAmerica/
Caribbean
Mathematics
1960 18.6(13)
1980 19.1(13)
Sciences
1960
1980
10.5(13)
11.6(13)
Sub-Saharan MiddleEast/
Asia
Africa NorthAfrica
OECD
Countries2
Totals
15.2(14)
17.9(14)
15.1(15)
17.5(15)
15.9(17)
17.0(17)
17.9(20)
19.3(20)
16.6(79)
18.2(79)
8.6 (14)
7.6 (14)
5.2 (15)
7.1 (15)
6.9 (17)
7.6 (17)
6.1 (20)
6.8 (20)
7.3 (79)
7.9 (79)
B. AverageAnnualHoursof Instructionin Mathematicsand Scienceat the PrimaryLevel
LadnAmerica/
Sub-Saharan MiddleEast/
Caribbean
Asia
Aflica NorthAfrica
Mathematics
1960 168.2 (7)
130.9 (11)
128.4 (7)
130.4(13)
1980 171.8 (7)
156.7 (11)
151.8 (7)
149.6(13)
Sciences
1960
1980
88.3 (7)
85.3 (7)
71.9 (11)
62.6(11)
54.8 (7)
66.0 (7)
53.2 (13)
68.1 (13)
OECD
Countries2
Totals
167.5 (17)
170.7 (17)
146.5 (55)
160.6 (55)
56.4 (17)
58.8 (17)
62.6 (55)
65.8 (55)
C. Emphasison Mathematicsand ScienceEducationat the LowerSecondaryLevel (1980sonly)
LatinAmerica/
Sub-Saharan MiddleEast/
Caribbean
Asia
Africa NorthAfrica
Mathematics
Percent3 14.5 (17)
14.9 (10)
16.5(18)
13.5 (17)
4
WeeklyHrs 12.5 (19)
12.1 (10)
13.3(19)
11.5 (17)
Sciences
Percent3 17.0(17)
WeeklyHrs4 13.6 (18)
17.8(11)
14.2 (11)
14.6 (18)
11.7(19)
13.3(16)
11.1 (17)
OECD
Countries2
Totals
14.0 (16)
11.8 (17)
14.7 (78)
12.3 (82)
16.7 (16)
14.0 (17)
15.7(78)
12.8 (82)
1.Unweightedaverages,numberof cases in parentheses
2. This categoryincludesWestemEurope,NorthAmerica,Australiaand New Zealand.
3. This row refers to thepercentageof the officiallowersecondarycurriculumdevotedto each subject
4. This row refers to the averagewekly hoursof instructionat the lowersecondarylevelin each subject
Table 2: Means and Standard Deviations for All Variables
in Panel Analyses by Level of Development
All
Countries
(n=63)
Log of per capita RGDP 1985
LogofpercapitaRGDP 1975
Log of per capitaRCDP 1960
bkmSD
3.40 .46
3.35 .42
3.15 .38
Mean YearlyInstructionalHours(circa 1960)in:
870.8 163.6
AllElementarySchoolSubjects
LanguageEducation
Mathematics
Sciences
History,Geog., SocialStudies, Civics
Arts,Music,Dance
PhysicalEducation
Religion,MoralEducation
Pre-Vocational,Practicalsubjects
PrimaryEnrollmentRatio 1960
SecondaryEnrollmentRatio 1960
TotalFertilityRate 1965
Dumny Variablefor Mining/OilExporters
ExportCommodityConcentration1960
305.6 109.3
144.5 40.4
63.3 42.3
75.6 36.5
81.2 36.3
58.7 30.3
50.5 43.3
65.4 69.2
77.4 27.3
24.2 19.0
4.9 1.9
.13 .34
46.4 30.2
More-DevelopedLess-Developed
Countries
Countries
(n=43)
(n=20)
M=a
3.89
3.80
3.56
=
.15
.15
.20
MM
3.18
3.13
2.96
S
.37
.34
.29
917.0 154.8
849.3 164.9
74.2
44.5
37.1
37.6
30.3
31.6
46.4
47.1
284.9 117.3
133.8 33.9
63.7 44.9
77.7 36.2
75.8 37.9
55.0 29.3
51.8 42.3
74.2 76.2
96.3 5.9
44.8 13.7
2.5 .44
.00 .00
17.6 15.5
68.7 29.0
14.6 12.5
6.0 1.1
.19 .39
59.8 25.6
350.0
167.5
62.4
71.1
92.9
66.6
47.6
46.4
Table 43: Correlation Matrix of Variables Used in Panel RegressionAnalysesa
(Sampleof Developedcountriesavocethe diagonal;sampleof LDCs belowthediagonal)
(1)
(2)
.93
(3)
.78
(4)
.84
(5)
.14
(6)
.37
(7)
-.01
(8)
.14
(9)
.20
(10)
-.02
(11)
.24
(12) (13) (14)
.04 -. 24 -. 81
(15) (16)
-.08 -. 66
.69
.82
.13
.39
-.06
.14
.12
-.02
.23
.05
-.07
1. RGDPpercapita
1985
2. RGDPper capita
.85
1960
3. Primay Education
.73 .61
1960
4. SecondaryEducaton
.77
.74
.60
-.04
-.09
-.05
.00
.22
.21
.37
.11
.42
.01
-.06
.06
-.10
.42
.66 .03 .39 .04 .01 .16 -.
05
.08
-. 17 -. 80
-. 58
.13 .13 -.22 -.63 -. 06 -. 39
.20
-.03
.04
.20
.08
.22
.02
-.09
.52
.36
.51
.20
.36
.50
.08
.26
-
14
-.19 -. 64
1960
5. Toal Instructional
Hours
6. Math Instructional
Hours
7. ScienceInstructional
Hours
8. Lang.Instructional
Hours
9. Arts Instructional
Hous
10. SocialSci. Instr.
Hours
11. PhysicalEd. Instr.
Hours
12. MoralRel. Instr.
Hours
13. VocationalInstr.
Hours
14.TotalFertility
Rate 1965
15.Dunmnyfor Mining/
OilCounuies
16.ExportCommodity
Concentration
-. 11
.22 .32 .00 -.
05
.35
-. 34
.14
.31
.37 -.
10 -. 07 -.39
.25
-. 10
.33
-.
15 -.
16 -.22
.46 .10 -.33
.06
-.09
.07
.04
.21
.02
.11
-.
16
.08
.13
.02
.20
.S3
.27
.46
-.16
.18
.14
.11
.31 .38 .22 .22 -.
07
.05
.04
-.15
.10
.05
-.06
-.13
-.17
-.05
-. 14
.14
.21
-. 18
.32
-. 44
-. 54
-. 53 -. 52
-. 72 -.07
-.21
.02
.12
-.09
-. 10
-. 29
.08
-.09
-.
13 -.
18
.23
.20
.10 -. 19
-.00
.16
.39 -.22 -.06
-.26
.15
-. 17
.02
.00
.09
-.14
.24
.03 .06 -. 28
.12
.03 .33 -. 01
.01
.08
.02 -. 25
-.07
.14 -.06
.00
.23
.16
.21
.07
.00
.12
.30
-.01
.16
.24
-.02
.09
-.10
-.12
.00
-. 35
-. 14
-. 12
-. 37
.04
.16
.22
-. 11
.00
.06
-.13
-.14
.06
.36
a) Zeo-order correiadonsin bold are significantat leastat the .05 level.
-.09
.07 -. 22
.12
-.21
.01
.05
.04 .03 -.36 -.
18
.44 .16
.07 -. 40
-.19
-.02
.19
.70
.38
.32
Table4
Effectsof CurricularContent and EducationalExpansion
on Economic Growth,1960-1985(Full sample)
(Unstandardizedregressioncoefficientsare reported; dependentvariableis per capitaRGDP 1985)
Equation:
(A)
AverageYearly
InstructionalTime (Total)
Average
Mathematics
Annual
Sciences
(B)
(C)
.00010
-.00021
.00102**
Instructional Languages
.00015
Timein
Ars/Music
.00064
Each
All SocialStudies
-.00001
Subject
Moral/Religion
.00059
Area
PhysicalEd.
-.00027
(circa 1960):
Pre-VocationalEd.
-.00042*
Educational
Primary
Education
.0043**
.0043**
.0042**
Rates:
Secondary
Education
.0034*
.0036*
.0035*
Control
Lagged
DependentVar.
.68**
.68**
.69**
Variables:
Mining/Oil
Exporters
.08*
.08
.07
Enrollnent
Export Commod.
Concentration
-.003**
-.003**
-.003**
Total Fertility
Rate (1965)
.015
.019
.019
AdjustedR2
.918
.918
.920
Constant
.91
.81
.74
Numberof Cases
63
63
63
* Unstandardizedregressioncoefficientis at least 1.5 timesits standarderror
** Unstandardizedregressioncoefficientis at least 2.0 times its standarderror
TableS:
Expansion and Curricular Content on EconomicGrowth,
Effects of Educational
(Less-developedcountries only)
1960-1985
(Unstandardizedregressioncoefficientsare reported; dependentvariableis per capitaRGDP 1985)
Equadon:
(A)
Mathematics
Annual
Science
-.00078
.00210**
.00015
Instructional Language
.00156*
Tunein
Arts/Music
Each
All SocialStudies
Curricular
Moral/Religion
Topic
PhysicalEd.
-.00180*
Atiune 1:
Pre-VocationalEd.
-.00094**
-.00055
.00083
.0043**
.0043**
.0031**
.0058*
.0057*
.0105**
.66**
.67**
.66**
Control
Lagged
DependentVar.
Variables:
Mining/Oil
Exporters
.10*
.10*
.07
Educational Primary
Education
EnrolLment
Secondary
Education
Rates:
Export Commod.
Concentration
*
**
(C)
.00006
Aver. Ann. Instruct.Time (Total)
Average
(B)
-.004**
-.004**
-.003**
TotalFertility
Rate (1965)
.044
.045
.034
AdjustedR2
.839
.835
.866
Constant
.79
.71
.82
Numberof Cases
43
43
43
Unstandardizedregressioncoefficientis at least 1.5times its standarderror
Unstandardizedregressioncoefficientis at least2.0 times its standarderror
Table6
InteractlonEffectsof CurricularContentand EducationalExpansion
on EconomicGrowth,1960-1985(Fullsample)
(Unstandardizedregressioncoefficientsare reported; dependentvariableis per capita RGDP1985)
Equation:
(A)
AverageYearly
InstructionalTime (Total)
Proportion
(B)
(C)
.00020*
Mathematics
-.00053
of School-age Sciences
.00125**
Cohort
Languages
.00044*
Exposedto
ArtsMusic
.00031
Instructional All SocialStudies
.00022
Hours in
MoraVReligion
.00055
SubjectArea
PhysicalEd.
-.00024
(circa 1960): Pre-VocationalEd.
Educational
-.00040*
Primary
Education
.0043**
.0027**
.0028*
Rates:
Secondary
Educadon
.0034*
.0037**
.0034**
Control
Lagged
DependentVar.
.68**
.67**
.70**
Variables:
Mining/Oil
Exporters
.08*
.08*
.08*
Export Commod.
Concentration
-.003**
-.003**
TotalFerdlity
Rate (1965)
.015
.021
.020
AdjustedR2
.918
.920
.920
Constant
.91
.90
.82
Numberof Cases
63
63
63
Enrollment
*
**
Unstandardizedregressioncoefficientis at least 1.5 timesits standarderror
Unstandardizedregressioncoefficientis at least 2.0 times its standarderror
-.003**
Table 7
Interaction Effects of Curricular Content and Educational Expansion
on Economic Growth, 1960-1985 (Less-developed country sample)
(Unstandardized regression coefficients are reported; dependent variable is per capita RGDP 1985)
Equation:
(A)
Average Yearly
InstructionalTime (Total)
(B)
(C)
.00022
Average
Mathematics
Annual
Sciences
.00280**
Instructional
Languages
.00030
Time in
Arts/Music
.00191*
Each
All Social Studies
Subject
Moral/Religion
Area
Physical Ed.
-.00198
(circa 1960):
Pre-Vocational Ed.
-.00133***
Educational
Primary
Education
.0043**
.0025
.0030
Rates:
Secondary
Education
.0034*
.0054**
.0122**
Lagged
Dependent Var.
.66**
Control
.684*
.67**
Mining/Oil
Exporters
.10*
.09*
.08
-.00148
-.00075
.00098
Enrollment
Variables:
Export Commod.
Concentration
*
**
-.004**
-.004**
-.003**
Total Fertility
Rate (1965)
.044
.048
.039
Adjusted R2
.839
.840
.867
Constant
.79
.71
.73
Number of Cases
43
43
43
Unstandardized regression coefficient is at least 1.5 times its standard error
Unstandardized regression coefficient is at least 2.0 times its standard error
PRE WorkingPaoerSeries
Contact
AuvhoDa
forpape
July 1991
S. King-Watson
31047
AndreaGubitz
WPS718 Impactof InvestmentPolicieson
GermanDirectInvestmentin Developing
Countries: An EmpiricalInvestigation
July 1991
S. King-Watson
31047
WPS719 HowTradeand EconomicPolicies
Affect Agriculture: A Frameworkfor
AnalysisAppliedto Tanzaniaand
Malawi
RamonLopez
RidwanAli
BjornLarsen
July 1991
M. Gunasekara
32261
WPS720 The Outlookfor CommercialBank
Lendingto Sub-SaharanAfrica
EllenJohnsonSirleaf
FrancisNyirjesy
July 1991
S. King-Watson
31047
WPS721 The Demandfor Moneyin Developing PatricioArrau
Countries: Assessingthe Role
Jos6 DeGregorio
CarmenReinhart
of FinancialInnovation
PeterWickham
July 1991
S. King-Watson
31047
WPS722 Is Rice Becomingan InferiorGood?
FoodDemandin the Philippines
MerlindaD. Ingoc
July 1991
P. Kokila
33716
WPS723 ImprovingWomen'sand Children's
Nutritionin Sub-SaharanAfrica:
An IssuesPaper
OlayinkaAbosede
Judith S. McGuire
July 1991
0. Nadora
31091
WPS724 Fiscal Issuesin Adjustment:
An Introduction
RiccardoFaini
Jaimede Melo
July 1991
D. Ballantyne
37947
WPS725 How Structureof Production
Determinesthe Demandfor Human
Capital
IndermitS. Gill
ShahidurR. Khandker
July 1991
A. Sloan
35108
WPS726 Perspectiveson the Designof
IntergovernmentalFiscalRelations
AnwarShah
July 1991
A. Bhalla
37699
WPS727 The Effectsof Debt Subsidieson
CorporateInvestmentBehavior
MansoorDailami
E. Han Kim
July 1991
A. Bruce-Konuah
80356
WPS728 Does Better Accessto Contracep- SusanCochrane
tives Increasetheir Use? Key Policy Laura Gibney
and MethodologicalIssues
July 1991
0. Nadora
31091
WPS729 Is Export Diversificationthe Best
Wayto AchieveExportGrowthand
Stability? A Look at ThreeAfrican
Countries
July 1991
M. Gunasekara
32260
WPS717 Does FinancialLiberalizationReally
ImprovePrivateInvestmentin
DevelopingCountries?
Jacques Morisset
RidwanAli
Jeffrey Alwang
PaulB. Siegel
PREWorkingPaperSeries
Contact
JjulQ
Author
for pape
WPS730 Wageand EmploymentPoliciesin
Czechoslovakia
Luis A. Riveros
July 1991
V. Charles
33651
WPS731 EfficiencyWageTheory,Labor
Markets,and Adjustment
Luis A. Riveros
LawrenceBouton
July 1991
V. Charles
33651
WPS732 StabilizationProgramsin Eastern
Europe: A ComparativeAnalysisof
the Polishand YugoslavPrograms
of 1990
FabrizioCoricelli
July 1991
Robertode RezendeRocha
R. Martin
39065
WPS733 The ConsultingProfessionin
SyedS. Kirmani
DevelopingCountries: A Strategyfor WarrenC. Baum
Development
July 1991
INUDR
33758
WPS734 CurricularContent,Educational
Expansion,and EconomicGrowth
July 1991
C. Cristobal
33640
Aaron Benavot