Taylor & Francis, Ltd. Bera

You might also like

Download as pdf or txt
Download as pdf or txt
You are on page 1of 25

BERA

Secondary Modern Schools: Are Their Pupils Disadvantaged? Author(s): Rosalind Levai and Alan J. Marsh Source: British Educational Research Journal, Vol. 33, No. 2 (Apr., 2007), pp. 155-178 Published by: Taylor & Francis, Ltd. on behalf of BERA Stable URL: http://www.jstor.org/stable/30032739 . Accessed: 12/09/2013 02:13
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp

.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.

Taylor & Francis, Ltd. and BERA are collaborating with JSTOR to digitize, preserve and extend access to British Educational Research Journal.

http://www.jstor.org

This content downloaded from 132.234.251.230 on Thu, 12 Sep 2013 02:13:39 AM All use subject to JSTOR Terms and Conditions

British Educational Research Journal 2007,pp. 155-178 Vol.33, No. 2, April

Routledge

Taylor & Francis Group

Secondary modern schools: are their pupils disadvantaged?


ofEducation, University ofLondon, alInstitute UK; bNottingham Local Education Authority, UK accepted 31 May 10 May 2004; resubmitted 25 January 2005; conditionally (Submitted 2005) 4 October 2005; accepted

Rosalind Levadic*aand Alan J.Marshb

(LEAs) that are whollyselectiveand a There are still 10 English local educational authorities and secondary further 10 withsome grammar modernschools.This articleexaminestheacademic performance of pupils in secondarymodern schools and the fundingof these schools using national data sets matchingpupils' performance at Key Stage 2 and General Certificate of Education (GCSE) as well as data on funding fromSection 52 statements. Studentsin secondary modern schools gained one less grade on average than equivalent studentsin comprehensive takingaccount of the cost school pupils obtainedfivegradesmore. After schools while grammar modernschools in the factors thatinfluence funding and grantentitlements per pupil, secondary years 2000/01-2002/03were funded around 80 less per pupil while grammarschool pupils schools. Secondary modern received over 100 more per pupil compared to comprehensive grammar deficit than comprehensive schools were more likely to be in financial and particularly are academicallydisadvantaged secondary modem schools, schools. Thus, students by attending to offset LEAs do not receivesufficient the depressing whichin most selective additionalfunding of the increasedsocial segregation effects on attainment arisingfroma selectivesystem.

Introduction
The deleterious effectof social disadvantage on student attainment is one of the key

from findings Student (PISA) surveys Assessment theProgramme forInternational of 2000 and 2003, which reportthe double impact of students'own family
background and the socio-economic status of their school's student body on attainment. The strengthof this relationship differsbetween countries. Some countries(Canada, Finland, Hong Kong-China, Iceland, Japan,Korea and Sweden) had above averageperformance in literacy and below average impact of social status on studentperformance (OrganisationforEconomic Cooperation and Development
of London, 20 BedfordWay, London, of Education,University *Corresponding author:Institute WC1H 0AL, UK. Email: r.levacic@ioe.ac.uk ISSN 0141-1926 (print)/ISSN1469-3518 (online)/07/020155-24 Educational ResearchAssociation 2007 British DOI: 10.2307/30000004

This content downloaded from 132.234.251.230 on Thu, 12 Sep 2013 02:13:39 AM All use subject to JSTOR Terms and Conditions

156 R. LevacicandA. J. Marsh findings [OECD], 2003). Therearesimilar for mathematics (OECD, 2004). Forboth literacy and mathematics there is also a group ofcountries aboveaverage surveys with Belgium, in schools(inparticular Germany socialsegregation is Hungary), which and associatedwith a stronger impactof social disadvantage on studentattainment. segregation socio-economic 'Amongthe ten countries withthe most pronounced into observedin PISA, all carryout selectionprocedures thatchannelstudents before or at theage ofassessment' ofsecondary education (OECD, different streams witha highdegreeof social segregation 2003 p. 220). In countries in schoolsthe literacy variance attainment schoolswas around60% ofthetotal in student between in student compared to 10% orlessin Scandinavian variance performance, countries, are muchless. schools'intakes wheresocialdifferences between thesocial variance in literacy scoreswas 21%, though The UK's between-school average students on of background The UK also impact attainment. had an above appears in PISA as a country secondary witha singletrack(i.e. non-selective) system, though one ofwhich, systems, actually thecountry consists offourdistinct selective England, Northern is wholly still has around20 local and another, Ireland, In thelight educationauthorities selective. (LEAs) whichare wholly or partially of 'where findings of sociothereis a highdegree segregation that along the PISA socio-economicbackgroundsdo economic lines, studentsfromdisadvantaged on 'secondary modem' schoolsin worse'(OECD, 2003, p. 223), thisarticle focuses EnglishLEAs, whichrecruit England.These are schoolsin selective theirintake test. As the current debate about school entry frompupils who fail a grammar standards, inequality educationand the comparative and diversity in secondary performance typesof secondary school gathers for pace (Department of different Educationand Skills[DfES], 2001, 2002a; Edwards& Tomlinson, for 2002; Office in Education[Ofsted], 2002; House of CommonsEducationand Skills Standards Committee, modernschools and theirpupils 2003), the positionof secondary deserves greater attention. and The purpose of this articleis to presentevidence on the performance resourcingof secondary modern schools in England and the consequent disadvantages incurred by youngpeople who attendthese schools. Most of the analysis is of data from nationaldata sets,but we also drawon a survey English of school costs undertakenas part of research conducted on behalf of the Upper SchoolsForum(Levacic etal., 2002), whichis campaignBuckinghamshire local authority ingforincreased funding forupper(i.e. secondary modern)schools. to the evidence of article academicperformance While on therelative adds existing and secondary grammar modem schools (Gallagher& Smith,2000; Schagen & there ifany,published has 2003), of been little, evidenceon thefunding Schagen, modem schoolsrelative types. secondary to other Our analysis of fallsintotwo mainparts.The first partis a value-added analysis
General Certificate ofSecondaryEducation (GCSE) results a nationaldata setfor from English secondary schools, with Key Stage 2 national tests as the measure of prior attainment. lower We findthatpupils at secondarymodern schools achieved slightly gradesat GCSE comparedto thosein comprehensive schools,whilepupils at grammar

This content downloaded from 132.234.251.230 on Thu, 12 Sep 2013 02:13:39 AM All use subject to JSTOR Terms and Conditions

pupils Aresecondary disadvantaged? 157 modern schools better. part The second didconsiderably examines the on average oftheanalysis secondary secondary ofEnglish schools. modem financing Despitesomeevidence that learning pupilshave greater schoolswerefunded less per modern needs,secondary schoolsin the three This yearsexamined. pupilthancomprehensive and grammar costsdue to variations in schools'size,region, age comparison allowsfordifferences rangeofstudents, need,suchas thepercentage ofstudents oflearning and indicators for eligible specialeducational needs. schoolmealsandwith free Background: selective education in England modernschoolshas been active The issue of selection forgrammar or secondary in 1945 of 'secondaryeducation for all' by the 1944 since the introduction such by a number contested EducationAct,and has been strongly oflobbygroups, as the National GrammarSchools Associationand the Socialist Educational Association. The steer byissuing comprehensive education was accelerated towards forEducationand Science, 1965) Circular10/65(Department the non-statutory which requested English and Welsh LEAs to provideplans for comprehensive education.Changes of government in the early1970s and the seriouseconomic education. slowedthedrive in A survey condition forcomprehensive ofthecountry 1975 indicated that only 20 of the English and Welsh LEAs were 'truly of all 10 year-olds comprehensive' stillsat the 11-plusexamination and a quarter (Simon, 1991, p. 439). In an attemptto acceleratethe process, the Labour Government thoseLEAs within Actto require passedlegislation the1976 Education thathad not responded to the 1965 Circular to do so. In 1979 a new Conservative was elected and repealed the 1976 Act. Over the next 18 years government comprehensive as a resultof redefinition educationwas subjectedto significant policiesdesigned to promote 'choice'and 'diversity', theselective with theeffect that todayto be wellentrenched system ofthe continues in 10-20 LEAs. The provision Act for a ballot of parentsto remove 1998 School Standardsand Framework schoolselection has not so fareffected anyfurther grammar change. In 2003 therewere 248 schools in England or 7% of all secondary schools' designated bytheAutumn as 'secondary PackageGCSE/GNVQ (General modern'2 of SecondaryEducation/General Certificate National Vocational Qualifications) (DfES, 2003a). Although the 'secondary benchmark information modern'schools 20 LEAs, only10 oftheseLEAs are classedbytheDfES as between are distributed wholly schoolballotbecause at selective, i.e. LEAs subjectto a possiblegrammar pupilsare in grammar schools.As Table 1 shows,in 2002 least 25% of secondary threeunder fiveof the whollyselective control, LEAs were under Conservative Labour and twounderLiberalDemocratcontrol.
A major factorin predisposing a selective systemto disadvantage childrenfrom less advantaged families,or who are at or below average in academic ability,is concentrating them in schools with otherless able and less advantaged children. In England only 2% of pupils in grammarschools are eligible for free school meals (FSM) compared to 14% in secondarymoderns. This is less than the average 18% of

This content downloaded from 132.234.251.230 on Thu, 12 Sep 2013 02:13:39 AM All use subject to JSTOR Terms and Conditions

158 R. Levacicand A. J. Marsh


Table 1. LEAs thatare definedby the DfES as whollyselective Selective LEAs LEA type Council control No. of No. of Percentageof No. of secondary secondary grammar pupils in grammar aged pupils modem schools schools schools 17,696 34,207 96,672 46,236 20,486 8,158 12,024 15,398 9,024 16,074 275,975 11 20 71 48 13 7 8 9 5 11 203 4 13 33 15 6 4 4 5 3 7 94 25 44 32 27 30 45 33 32 31 30 33

Bexley London BuckinghamshireUpper Tier Upper Tier Kent Upper Tier Lincolnshire Medway Unitary Slough Unitary Southend Unitary Sutton London Unitary Torbay Metropolitan Trafford Total

Lab Con Con Con Con Lab Con Lib Dem Lib Dem Lab (min)

Sources: DfES 2001 AnnualSchoolsCensusdataset, and DfES (2003a) Pupil LevelAnnualSchools Census.

pupils eligibleforfreeschool meals in comprehensive schoolsbecause almostall selectiveLEAs are in areas of above averagesocio-economic status.As Table 2 FSM in secondary modern schoolsin the for shows,thepercentage ofpupilseligible 10 selective to 11% in Buckinghamshire LEAs rangesfrom and 26% in Trafford schools, 5% in Sloughto 1% in Buckinghamshire Suttonand,for grammar from and Sutton.OnlyTrafford percentage ofpupilseligible has a higher forFSM thanthe LEAs. averagein non-selective A higherproportion of studentsin secondarymodern schools have special educational Table 3, 3.7% ofpupilsin secondary needs (SEN). As can be seenfrom
by school typeofsociallydisadvantaged for Table 2. Distribution pupilsas measuredby eligibility freeschool meals: Englishsecondaryschools (January 2001) SelectiveLEAs Bexley Buckinghamshire Kent Lincolnshire Medway Slough Southend Sutton Torbay Trafford All selectiveLEAs LEAs Non-selective LEA 13.2 7.1 10.2 10.1 11.7 17.2 12.2 7.6 13.1 20.3 12.3 19.8 Secondarymodem 16.5 10.9 14.2 12.6 15.4 24.2 18.0 11.1 18.6 25.7 14.3 Grammar 3.5 1.2 2.2 1.9 3.1 5.1 2.4 1.2 3.8 4.0 2.4

Source: DfES 2001 Annual SchoolsCensusdataset.

This content downloaded from 132.234.251.230 on Thu, 12 Sep 2013 02:13:39 AM All use subject to JSTOR Terms and Conditions

Aresecondary pupils disadvantaged? modern 159 modern to 0.1% in grammar schoolshad statements schoolsand ofSEN compared of pupils with SEN but schools,while the percentages 2.7% in comprehensive without statements of the was 24.4, 3 and 18.6% respectively. Another indicator additional needsofsecondary modern educational pupilsand thecoststheseimpose wereconsiderably Permanent on schoolsis therateofpupil exclusions. exclusions higherfor secondary modernschools (2.5 per 1000 pupils) and comprehensive schools(2.3 per 1000 pupils)thanforgrammar schools(0.3 per 1000 pupils). is associated demonstrate, Thus, thefacts with thatselection as one wouldexpect, disadvantaged of socially and educationally pupils in increased concentration schools-the secondary moderns.Since the socio-economic and prior particular of the peer group(the 'compositional effect') influence attainment characteristics individualchildren's academic attainment, a key issue in assessingthe relative of selective and non-selective schoolsystems the narrower performance is whether educational modern of secondary compared mission to comprehensive schoolscan student impactofa moredisadvantaged thenegative body.It is thisissuethat offset evidenceon therelative performance theexisting we focusupon in summarising of secondary modern and grammar schools. The academic performance of secondary modern and grammar school pupils: existingevidence A rational assessment and disadvantages ofselective oftherelative advantages school restson weighing and equitycompared systems efficiency to up evidenceon their inherently systems. school system increasessocial non-selective Since a selective efficiency segregation has to be that of greater in schools,its main justification (Gallagher 2000). A selective is moreefficient thana non-selective & Smith, system on average becausethey in schoolsthat one ifstudents attain morehighly aretaught focuson a narrower have needsthannon-selective rangeoflearning schools, which to servethe wholeability be than A more efficient a could range. selective system one evenifsecondary modern pupilsattained lessthanthey wouldin comprehensive system. for thelatter This wouldbe thecase ifpupilsselected grammar schoolsgain higher the compared to outweigh sufficiently system attainment to a comprehensive pupilsso thatselective forsecondary modern produce reducedattainment systems systems.However, if betteracademic resultson average than comprehensive
pupils: forEnglish Table 3. Pupilswith by schooltype specialeducational needsand excluded (January schools 2001) secondary Schooltype ofpupilsPercentage Percentage SEN Permanent exclusions with ofpupils statements butwithout % statements % per 1000pupils with 3.7 0.1 2.7 2.7 24.4 3.0 18.6 18.3 2.5 0.3 2.3 2.1

Secondary modern Grammar Comprehensive England: all schools

Sources: Annual Schools Census (2001) and Ofsted(2002).

This content downloaded from 132.234.251.230 on Thu, 12 Sep 2013 02:13:39 AM All use subject to JSTOR Terms and Conditions

160 R. LevacidandA. J. Marsh ability pupilsperform worsethanthoseof equivalent secondary and social modern background is clearly in comprehensive inequitable. schoolsthena selective system therefore research seeksto provideevidenceon the efficiency Policy-relevant and impacts systems distributional of schooling. ofselective of different schools must controlfor A valid comparisonof the performance is priorattainment performance students' to theschoolswhoserelative before entry then assessed in termsof the progresspupils make betweengiven phases of education.Two main typesof data have been analysedin this research:crossand non-selective yearin selective sectionaldata of pupil cohortsin a particular data fromcohortstudiesof people from and longitudinal schools and systems, childhood to adulthood. A brief review of previous researchon the impact of selection on pupil by Schagen and Schagen (2003) in thisjournalincludedan earlier performance of theliterature papersbyJesson unpublished review by Crooketal. (1999), three Irelandby (1999, 2000, 2001) and a studyof GCSE performance in Northern ofCrooketal.'s review Shuttleworth conclusion and Daly (2000). The tentative was werecontrolled difference was little variables forthere thatonce pupilbackground overallresults. and non-selective betweenselective systems' Jesson'sanalysis used measure Key Stage3 as theprior better forGCSE and reported slightly attainment systems.In contrast, Shuttleworth performance and Daly's by comprehensive which restricted to comparing and secondary (2000) study, was necessarily grammar datafrom Ireland, found thata student modemschoolsonlyas itanalysed Northern at of givensocial background school obtained16 and priorattainment a grammar thesame characteristics to a pupilwith morepointstotalscoreat GCSE compared modernschool. at a secondary foundno conclusive evidence forsignificant differences A DfES (2002a) study in pupil progress in LEAs classifiedinto three selection types-no selection, The exceptions selection and selective. werepupilswithKey Stage 2 intermediate attainment oflevel4 or above,whomadethemostprogress overKey Stage3 in prior selective LEAs LEAs, and at Key Stage 4 wherepupilsin schoolsin non-selective GCSE more gained 1-2 LEAs. points thanin selective did notuse multilevel unliketwostudies by The DfES (2002a) report modelling, researchers fromthe National Foundationfor Educational Research (NFER). Schagen and Schagen (2003), using Qualifications and Curriculum Authority foundthata pupil withan (QCA) nationaldata setsforaround400,000 pupils,3 halfa levelmoreprogress to Key scoreof4.5 at Key Stage2 made roughly average Stage3 in a grammar school.More able pupilswith schoolthanin a comprehensive in a grammar level5 orhigher school.Therewas only at Key Stage2 did little better a smalldifference schoolsfrom ofgrammar Key Stage3 to GCSE, and no in favour Selective for measured difference pupilsofhighest ability. systems, as thepercentage
slightly of pupils selected in an LEA, performed betterat Key Stage 3 because of the impact of grammarschool attendanceon the least able of the grammarschool pupils. large A latermultilevelstudyby Benton et al. (2003)" found a positiveand relatively fromKey Stage 2 to Key Stage 3 and fromKey Stage 2 to grammarschool effect

This content downloaded from 132.234.251.230 on Thu, 12 Sep 2013 02:13:39 AM All use subject to JSTOR Terms and Conditions

pupils disadvantaged? 161 modern Aresecondary pupilsin grammar GCSE. The lowestability schoolsachieved7 GCSE totalscore points(grades)morethantheywould have done in a comprehensive school. The grammar and was non-existent forhighschool advantagedeclinedwithability pupils (thosewithlevel5 and above at Key Stage 2). ability have the effects A numberof studiesreporting of selectiveschool attendance analyseddata from the National Child Development Study (NCDS) of children for the born in the same week in March 1958. Steedman (1980), controlling no statistically characteristics background ofpupils,reported differences significant in pupils' progress In a laterstudy and comprehensive in selective schoolsystems. that grammar invalueaddedbetween Steedman(1983) found difference theaverage and comprehensive school pupils was 0.92 O level grades and that between comprehensive and secondary modern schools0.28 grades. of grammar Later studiesusingNCDS data have reported on the effects school for familybackgroundand cognitive attendanceon qualifications controlling attainment at the ages of 7 and/or 10. Feinsteinand Symons(1999) foundthat a attending grammar of school added 7.6 points to O level/CSE (Certificate SecondaryEducation) scoreswhilebeing at a secondary modem cut themby 2 school). Dustmann et al. (2003) reportedthat (compared to comprehensive school attendance was associatedwithtwo more O level subjectpasses grammar comparedto beingat a secondary modernschool.The main purposeof Galindobenefited Rueda's and Vignoles'(2004) study was to testwhether individuals from for areasin the1970s,controlling ability at age 7, a range beingeducatedin selective indicators for area of residence.Four of familyvariablesand socio-economic measures were at age 16, years ofschooling, employed-maths score Aofoutcome Grammarschool attendanceresultedin higher level and highestqualification. whilelow or averageability outcomesforall ability groups, students in secondary no There to comprehensives. appearedto be no effect did worsecompared moderns area but there on averageoutcomesofbeingeducatedin a selective was a positive effect ability quintile.5 forthehighest generally thatgrammar studies and theNCDS studies find Boththeschoolcohort school attendance school improves academicresultscomparedto comprehensive them, attendance though and secondary modern thelatter result schoolsworsen is moreequivocal.The school cohortstudiesindicatethatit is middleability pupils mostfrom grammar schooland do relatively worsein comprehensives. who benefit do notreport The NCDS studies on the They are,however, thisfinding. reporting comprehensive effect ofa well-established created system relative to a newly selective in the 1970s-whereas verydifferent schools conditions system in comprehensive by thelate 1990s, in particular themuchgreater attached to prevailed importance examination performance.
Comparison of pupil performance from Key Stage 2 to GCSE in secondary Our analysisis similarto Benton etal.'s but uses the first large-scalenational data set at pupil level provided by the QCA for GCSEs sat in 2001 matched with prior

modern, grammar and comprehensive schools

This content downloaded from 132.234.251.230 on Thu, 12 Sep 2013 02:13:39 AM All use subject to JSTOR Terms and Conditions

162 R. Levalicic andA. J. Marsh also employs attainment Key Stage2 (KS2) in 1996. Our analysis datafor multilevel The advantage of ofthismethod modelling. is thatit allowsforthecommoneffect the statistical of its pupils and thus removes each school on the attainment bias in a statistical modelwhich present ignores that thefact pupilsarenestedin schools6 1995; Kreft 2003, p. 580). The & De Leeuw, 1998; Schagen& Schagen, (Goldstein, QCA matcheddata set includesthe National Curriculum test and examination results, as well as age, genderand school7of over 330,000 stateschool pupils.A rangeof school context of the Register variables(listedbelow) was obtainedfrom for2000 and from EducationalEstablishments theAnnualSchools Census forthe years1997 to 2001.8 ofpupils'GCSE/GNVQ attainment areused:(i) GCSE/GNVQ total Two measures as usualbygiving score(derived points an A* grade8 points, A=7, B= 6, C=5, D=4, E=3, F=2 and G= 1 and addingup the gradesthe pupil obtained);and (ii) the A*-C GCSE/GNVQpasses.The explanatory probability that a pupilobtains 5 ormore ofGCSE/GNVQ attainment attainment for bothmeasures variables areprior at KS2 for age and gender. (average mathematics and science),9 English, marks At schoollevelwe includedthepercentage forFSM, with ofgirls, pupilseligible ratioto take and who were white,as well as the pupil-teacher SEN statements accountof differences in resourcing. All theschool-level continuous variables were the context of the school averagedoverthe fiveyears1996/97-2000/01 to reflect Year 7 to Year 11. Descriptive whilethepupilswereprogressing from statistics are givenof the main continuous set (Appendix). Table Al variables A of dummy in was included highest type variables ofdenominational schooland for age being16,10 specialistor secondary the school was a grammar, whether modernschool (the comprehensive school witha sixth school being a non-denominational reference A dummy term selective LEAs was includedas a testofwhether form). forwholly beingin a selective on GCSE attainment LEA had an additional effect in addition to thatofthetypeofselective schoolattended. in MLwiN The regressions werefitted at three forvariance levels-LEA, schooland pupil. allowing The effect modernor a grammar school is testedby a secondary of attending on average,the GCSE resultsof pupils in thesetwo typesof selective whether, thanthosepredicted forpupilswiththe same KS2 test schoolwereless or greater comprehensive results schools.For the purposesof thistestwe are who attended onlyconcernedwithhow well KS2 testsprovidea measureof pupils' cognitive pupilsand which attainment oftheir whichis consistent is a good predictor between laterGCSE performance. issue of whether or not The much more controversial key increasing educational standards stage scores rising test over time mean 2004; Statistics 2005) is notrelevant ofa single Commission, fora study (Tymms, cohort ofpupils.In fact, Tymms(2004) reports thattheKS2 testsformathematics and Englishcorrelate testsforYear 6 pupils.11 A well withthe PIPS attainment
further indicationof the validityof the KS2 testsis thatin our data set theycorrelate well with teacher assessment in the respective core subjects (for English the lower at 0.76). Our main correlationis 0.82, mathematics0.83 and science slightly concern is withthe abilityof the KS2 total score in English,mathematicsand science

This content downloaded from 132.234.251.230 on Thu, 12 Sep 2013 02:13:39 AM All use subject to JSTOR Terms and Conditions

163 Aresecondary modern pupils disadvantaged? to predict GCSE results and so provide a validmeasure for comparing pupilprogress modernschools.Fifty-two grammar and secondary in comprehensive, per cent of in pupils'GCSE scoresis explained bytheir thevariation KS2 totalscore,whichis highly significant. a multilevel model including gender,age and the school context From fitting effect of attending variableslistedabove, we foundthatthe averageestimated a KS2 scores)was 5.5 grammar (as measured schoolon pupilsofall abilities bytheir additional to beingat a comprehensive GCSE/GNVQ points(i.e. grades)compared of attending a secondary modern school. The effect was on average1 gradeless at GCSE comparedto a comprehensive school. These resultsare givenin the first estimated columnofTable 4, whichalso contains coefficients on a selection ofother equation.12 are independent (The fullregression variables in the regression results differin Table A2 in the Appendix.)The grammar-secondary-modern reported is considerably ential reported byShuttleworth at 5 grades smaller thanthe16 points Ireland. and Daly (2000) forNorthern the effect We testedwhether on attainment varies ofthetypeof schoolattended to ability terms betweenthe pupil's KS2 score interaction according by including a grammar school,and between and attending theKS2 scoreand secondary modern coefficient on the grammar school. The estimated school-KS2 score interaction was negative, that themoreable thepupil,thesmaller indicating term theadvantage schoolcompared to beingat a comprehensive. ofattending A pupilwith a grammar to gain9.4 GCSE pointsby attending a mean KS2 scorewas estimated a grammar school comparedto being at a comprehensive, whereasa pupil with 1 standard abovetheKS2 meanonlygained4 additional GCSE points bybeingat a deviation grammar The results thatthemoreable thepupil (at KS2), the school.13 indicate greaterthe loss in GCSE points fromattending a secondarymodern school. an obtained one average student ability Whereas less GCSE gradeat a secondary
Table 4. Estimates ofeffect on GCSE totalscores ofschooltype Independent variable Size ofeffect on GCSE total scoreofa 1 unit variable change in theindependent interaction Without terms KS2 totalscore Girl free pupils eligible for schoolmeals Percentage school Grammar modern school Secondary Wholly selective LEA KS2 scoreinteracted with grammar school KS2 scoreinteracted with secondary modern school 0.323 3.375 -0.146 5.50 -1.04 0.27 (notsig) omitted omitted interactions With 0.328 3.375 -0.146 9.37 -1.23 0.02 (notsig) -4.94 -1.04

Note:all coefficients arestatistically significant at .05 unlessotherwise stated.

This content downloaded from 132.234.251.230 on Thu, 12 Sep 2013 02:13:39 AM All use subject to JSTOR Terms and Conditions

andA. J. Marsh 164 R. LevalicW KS2 scoreabovethemeanwouldachieve modem,a pupilwith1 standard deviation 2 gradesless. ofa pupilobtaining 5 The secondmeasure is theprobability ofGCSE attainment or moregradesat A*-C. This requires fitting regression equation(briefly a logistic results bySchagenand Schagen, inTable 5.14 explained 2003). The aresummarised A girlor boy of averageability 5 or more of getting had a 30% higher probability a comprehensive a grammar schoolthanifattending gradeA*-C GCSEs attending modern. 3 and 4% lesschanceifat a secondary For pupilsofhigh and onlybetween (1 standard deviation to achieve5 ability above average) boyswere15% morelikely 10% in a girls more and both likely, good GCSEs grammar schooland were2-3% of forbothmeasures less likely to achievethisin a secondary modern.Our results GCSE attainment-that school middle abilitychildrengain most by grammar attendance-areconsistent withthoseof Schagenand Schagen(2003) and Benton etal. (2003). schoolson attainment to whichthe effect of grammar The extent is due to the positiveinfluence and well-motivated peer groupand to of a sociallyadvantaged due to theseschools'focused is difficult to determine whatextent academicmission & Daly, 2000). Estimates of the grammar-secondary (Shuttleworth modern attainment which use a dummyvariableto measure the effect differential on of disadvantage attainment of the two school typesunderestimate the attainment pupilbecausethemeasurable beinga secondary effects modern ofthepeergroupare byincluding ofpupilseligible forFSM orwith foralready thepercentage controlled To get idea of secondary a better oftheattainment we moderns SEN. disadvantage oftheeffect schools on pupils'attainment ofattending need to includean estimate disadvantaged students ofsocially would concentration thathave a higher thanthey in mostcases ifthelocal system werecomprehensive. selective In the 10 wholly ofpupilseligible LEAs theaverage percentage forFSM in 2.3% compared FSM secondary average modems.The percentage was to 14.8% schools for therefore, theLEAs as a wholewas 10.7. On average, secondary modern
Table 5. Estimates modern on theprobability of andsecondary ofgrammar oftheeffect schools at GCSE compared achieving school A*-C grades 5 or more to a comprehensive ofa student Increased probability gaining 5+A*-Cgrades at GCSE in terms ability ofKS2 scores Ifpupilofaverage deviation Ifpupil1 standard aboveaverage in terms ability ofKS 2 scores Ifpupil2 standard aboveaverage deviation2 ofKS 2 scores ability in terms Boy Girl Boy Girl Boy Girl modern Grammar Secondary school% school% 30.0 29.5 14.8 10.0 2.8 1.7 -3.4 -4.0 -3.0 -2.2 -0.7 -0.4

Sources: QCA matched SchoolCensus2001. pupildataset 1996-2001;Annual average ability ofKS2 results); 34% (interms pupils were atorbelow school Note:3% ofgrammar deviations were between 3.1% were aboveaverage 2 standard ability; 0 and 1 standard deviation or ability. moreaboveaverage

This content downloaded from 132.234.251.230 on Thu, 12 Sep 2013 02:13:39 AM All use subject to JSTOR Terms and Conditions

Aresecondary modern disadvantaged? pupils 165 system. FSM due to theselective had 4% morepupilseligible for Usingtheestimate oftheeffect FSM on GCSE totalpointsof -0.146, theeffect of 1 percentage ofthe is to reduce attainment in secondary modernson 4% higherFSM percentage schoolpupilsenjoytheadvantage by 0.6 GCSE grades.Grammar average of 8.4% on oftheir peergroupentitled to FSM, which fewer positive has an estimated impact GCSE scoresof 1.2. The advantages peer groupforthoseselectedfor of a better estimated to be 1.8 additional grammar schoolare therefore grades. caused impact ofsocialstratification As thePISA reports pointout,ifthenegative selective by selective by dismantling systems, schoolsystems is not to be addressed then to raisetheattainment ofsocially policies pupilsneed 'to improve disadvantaged the resources, policies,processesand climate'in the schoolswhichthesestudents thefunding ofsecondary attend(OECD, 2003 p. 267). However, whenwe examine per modernschoolsin Englandwe findthatthey are in factless generously funded after schools, taking which offactors and grammar account pupilthancomprehensive determine cost differentials, schoolcosts such as size, age range,regional and the pupilsand thosewithspecialneeds. disadvantaged ofsocially concentration The fundingof secondary modern schools Data on the publicrevenues by Englishsecondary schoolswereobtained received fromthe Section52 outturn statements on individual school budgetsthatLEAs to theDfES. The Section52 statements annually submit (DfES, 2002b) provide two measures ofrevenue per pupilper annum: allocated * budgetshare perpupil(therevenues plusthose byLEA funding formula forms); whichcome via theLearning and SkillsCouncilforsixth * current devolved revenueper pupil, whichincludesin additioncontingency, fundand other central grants. standards analysisof the data and then a regression We firstpresentsome descriptive determinants includetheschool-level variables ofschoolfunding. The data analysed used in the analysis of attainment Section52 statements. plus thoseadded in from The schools' budget sharesper pupil15and current revenueper pupil for the schooltypes years 2000/01-2002/03 in Table 6. financial thethree are compared for as thesereceive are compared sixth forms morerevenue Onlyschoolswith perpupil and secondary without have disproportionately more schools sixth moderns forms. arehigher ofsocially concentrations and SEN pupilsin Thoughthere disadvantaged secondary modernschools,therewas littledifference in budgetshare per pupil in favour ofsecondary schools, though thismovedslightly withgrammar compared The funding modernsoverthe threeyears2000/01-2002/3. per pupil was more

favourableto secondarymodernswhen the amounts receivedvia centralgovernment are included. The patternis also different by local authority. We show this just for 2000/01 in Table 7. In seven of these LEAs in 2000/01 the budget share per secondarymodern pupil was less than the grammar school budget share per pupil. Slough was the most

This content downloaded from 132.234.251.230 on Thu, 12 Sep 2013 02:13:39 AM All use subject to JSTOR Terms and Conditions

andA. J. Marsh 166 R. Levacidc


Table 6. Fundingof secondaryschools withsixthforms by school type Budget Share per pupil Total public revenueper pupil 2002/03 3258 3151 3342

2000/01 2001/02 2002/03 2000/01 2001/02 Secondarymodem schools Grammarschools schools Comprehensive 2507 2561 2627 2703 2694 2767 2963 2936 3028 2675 2652 2804 2934 2844 3037

Source: Section 52 Budget Statements, 2000/01 to 2002/03 (DfES, 2002b).

generouswith 330 extra per secondarymodem pupil, but it had only one secondary form. a sixth modern schoolwith Sutton was thenextmostgenerous with modernpupil, while Kent provided47 extra.Two ,196 extraper secondary Buckinghamshire and Torbay, funded each grammarschool pupil authorities, around90 more. However, thecomparisons in Tables 6 and 7 do nottakeintoaccountdifferences thatdetermine in the factors school costsper pupil-that is, the cost drivers. The maincostdrivers schoolsize),theage costsfall arethesizeofschool(as average with rangeof pupils (since costsper pupil risethrough keystagesas the the secondary bythe curriculum in regional costs(proxied becomesmorespecialised), differences needs,whichareproxied AreaCost Adjustment Factor16 learning by and additional ofpupilseligible forFSM and withspecialneeds. indicators such as thepercentage
schoolsin selective LEAs Table 7. Budgetshareperpupilin secondary modemand grammar
2000/01 Schools withsixthforms All schools

Secondary Grammar Difference between between Secondary Difference moderns Schools secondarymodern moderns secondarymodem and grammar withsixth and grammar funding forms funding Bexley Buckinghamshire Kent Lincolnshire Medway Slough Southend Sutton Torbay Trafford 2500 2253 2529 2448 2507 2978 2663 2610 2438 2421 2530 2341 2482 2498 2538 2648 2713 2414 2532 2464 -30 -88 47 -50 -31 330 -50 196 -94 -43 2524 2253 2539 2594 2496 2864 2759 2594 2433 2671 -6 -88 57 96 -42 216 46 180 -99 207

2000/01 (DfES, 2002b). Source: Section 52 Budget Statements since it had extremely is omitted Note: in Medway one secondary modernwithonly 128 students highcosts due to its size.

This content downloaded from 132.234.251.230 on Thu, 12 Sep 2013 02:13:39 AM All use subject to JSTOR Terms and Conditions

disadvantaged? 167 modern pupils Aresecondary In additionto the LEA formula-determined budget,schools receiveddifferential amounts Fund via theDfES, someofwhich was due to participating ofStandards in startand particular such as Excellencein Cities,and beacon, fresh programmes and in in cost drivers specialistschools.We need to allow forthese differences foradditional eligibility secondary StandardsFund beforewe can judge whether modern and grammarschools are funded more or less than comparable comprehensive schools. This is done by regressing pupil17 forthe years current per secondary revenue 2000/01-2002/03 listedabove and on a set of dummy on the cost drivers variables categories grammar ofschool for and secondary thedifferent schooland for modern interest standards inTable 8. Ofparticular The results aregiven receiving funding.'8 on thedummies significance is thesize and statistical ofthecoefficients grammar for or ofhowmuchextra school.These showtheestimate schooland secondary modern less a schoolgotperpupilifitwas a grammar modern compared to a ora secondary school aftercontrolling school comprehensive for otherfactorsthat determine are consistently positivefor grammar revenues.These coefficients schools and significant statistically in all threeyears.Grammar schoolswerefundedmoreper 115 in 2001/02and pupilcompared to comprehensive schools(131 in 2000/01, 204 in 2002/3)after allowing in school-level fordifferences Standards costdrivers, Fund eligibility, differences in LEAs' estimatedneed to spend on secondary education(measured by StandardSpendingAssessment or SSA) and forthe fact selective selective LEAs fundless per pupil. On averagewholly LEAs thatwholly schoolsperpupil,varying from 56 in 2000/01 spentless on secondary an estimated modernschoolswerefunded forthis,secondary to 83 in 2002/03.Even allowing of schools.The estimated coefficients less per pupil comparedto comprehensive for 2000/01 and 2001/02 but the -85 and -73 are statistically significant amountfor2002/03thoughstillnegative is not statistically significant. Grammar are estimated to have obtained200 moreper pupilin thatyear schools,however, thana comprehensive cost drivers. schoolwithequivalent threeyearsof school revenuedata therefore shows thatsecondary Examining schools disadvantage compared modern to comprehensive have been at a financial schoolshave receivedrelatively schools in two of the yearswhilegrammar more generous all three funding years.19 Budget deficitsand surpluses oftherelative ofsecondary A further underfunding indication modemschoolsis that schools,while than comprehensive theyare more prone to run budget deficits be grammar in the red. Section 52 statements schoolsare in much less likely to includethebalancecarried to thenextfinancial forward year.Fromthistheschools carry in each yearfrom2000/01to 2002/03 forwards (i.e. deficits) withnegative zeroto three wereidentified ofyears from each schoolwas in deficit and thenumber by school type.The distribution of schoolsby was calculatedand cross-tabulated to number thatdeficits ofyears wererunis shownin Table 9. Fiftytypeaccording per centofsecondary werein surplus yearscompared four to moderns forall three

This content downloaded from 132.234.251.230 on Thu, 12 Sep 2013 02:13:39 AM All use subject to JSTOR Terms and Conditions

168 R. LevacicandA. J. Marsh


Table 8. Variables explainingpatternsin revenueper pupil for secondaryschools 2000/01 to 2002/03 Independentvariable 2000/01 2001/02 1895** 0.107* 383038** 14.16** 28.65** 1.55** 133.23** 115.19** -73.00** -67.02** 43.38** 70.33* -59.81** 17.95 39.81 393.74** 47.34** 67.14** 89.48** 2532 0.72 2002/03 1987** 0.145** 422703** 17.21** 38.69** 2.87** 154.47** 204.39** -43.46 - 83.46** 69.19** 144.0** -38.52 34.06** 94.34** 407.54** 67.12** 48.99** 62.43** 2750 0.74

Constant 1838** No. of pupils enrolled 0.080** Inverseof no. of pupils enrolled 306483** Percentageof pupils eligibleforfree 12.05** school meals Percentageofpupilswithstatements 23.37** of SEN Percentageof pupils withSEN but 1.09 withoutstatements 128.73** School has a sixthform 131** Grammarschool -85.23** Secondarymodernschool -55.50* School in whollyselectiveLEA Specialistschool 50.3** 40.57 School in special measures School in education actionzone -34.72 Beacon school 5.2 Excellence in Cities school 50.05* Fresh startschool 531.22** Trainingschool 27.27 31.36* Leading edge school grantschool 53.96** Leadership incentive 2452 No of schools Adjusted R squared 0.69

at .05 and * at 0.1. Note 1: ** indicatesstatistical significance of outliers and clustering forby ofschoolsin LEAs on standarderrors Note 2: the effect corrected using robustand clustercommandsin Stata. 64% of comprehensives. Fifteen per cent were in deficit for all three years compared to 9.4% of comprehensives. In contrast, 88% of grammar schools were in surplus in

differences between thedistribution all three for years. The chisquarestatistic ofthe schooltypes three was significant by number ofyearsin deficit at 1%.20 severein The deficitpositionof secondarymodem schools was particularly Buckinghamshire 9 out of20 schoolshad rundeficits three where for yearsand only surplus 1 had in had run every grammar schools 13 in Only the year. outof four been throughout theperiod. forthree yearswhile11 had been in surplus a deficit have a numberof deleterious implications forschoolsand their Budgetdeficits A school witha budgetdeficit cannotapplyforspecialist students. schoolstatus, policy advocates as a means for improving the quality of which government
secondaryeducation and which provides a school with additional funding.A school with a deficit has no contingency fund to use in the event of unexpected

expenditures.A deficitmust be eventuallycorrectedand, unless additional funding is provided, this can only be achieved by spending less than currentrevenue, thus

This content downloaded from 132.234.251.230 on Thu, 12 Sep 2013 02:13:39 AM All use subject to JSTOR Terms and Conditions

pupils disadvantaged? Aresecondary modern 169


bytype deficits 2000/01-20002/03 Table 9. Percentage schools ofsecondary budget with in budget 2000/01-2002/03 Number in period ofyears deficit Zeroyears Secondary modern Number ofschools % Grammar Number ofschools % Comprehensive Number ofschools % total/average England ofschools Number % 129 54.4 131 87.9 1586 64.3 1846 64.7 1 year 41 17.3 11 7.4 362 14.7 414 14.5 2 years 32 13.5 3 2.0 285 11.6 320 11.2 3 years 35 14.8 4 2.7 232 9.4 271 9.5

52 Statement (DfES, 2001b,2002b,2003b). 2001/02, Source:Section 2002/03 2000/01, 2001/02 for datafor and 2002/03. Note:no financial Trafford

the educationof the current impoverishing cohortof students in orderto pay for cohort received. whata previous health ofschoolspoints The foregoing evidence on funding and thefinancial quite to secondary bylower clearly modempupilsbeingon average disadvantaged funding schools, and particularly perpupilthanpupilsin comprehensive relative to grammar schools. The additional costs of secondary modern schools Upper Budget deficitproblemsspurredthe formation of the Buckinghamshire Schools Forum to campaignfor what theyregardas fairerfunding of upper to estimate theadditional costs In order modem) schoolsin thecounty. (secondary modemschoolsincurcompared to grammar thatsecondary schoolsbecause ofthe greater needs of their student communities,the Forum commissioneda schools.21 From initial to be administered to its member discussion questionnaire additional with head teachers and questions askedabout costfactors wereidentified to staff thesein relation timeinvolved and otherexpenses. the questionnaire on the costs of Thirteenof the 21 upper schools completed aspects of provision, estimated different and the additional costsare an averageof foradditional as partof a campaign these.Giventhepurposeof the questionnaire

resources,one would expect head teachersto have an incentiveto bias upwards the costs reported.We attemptedto counteractthisby makingconservative assumptions when deriving the cost estimatesand by not includingexpenditures thatwere funded fromearmarkedrevenues.Nevertheless, we cannot be certainthatan upward bias in the estimatedcosts is not present,thoughthe argumentthat secondaryschools have

This content downloaded from 132.234.251.230 on Thu, 12 Sep 2013 02:13:39 AM All use subject to JSTOR Terms and Conditions

170 R. Levac'liandA. J. Marsh learning higher needsoftheir coststhangrammar schoolsdue to thegreater students is highlyplausible. The specificneed-related activities that were included as below and the estimated cost per costs are explained additional briefly additional student are givenin Table 10. Vocational courses These are offered proportion in years10-12 in upperschools to a higher ofstudents compared to grammar to run. schoolsand are usually morecostly Specialeducational needs The percentage of SEN pupilswithand without in secondary a statement modem schools far outweighs SEN expenditure that in grammar schools. Therefore in in secondary grammar schools can be regarded as verylow, so thatexpenditure modernschoolson SEN is treated as additional costs.The number and ofteachers students support in supporting staff to be deployed withSEN was costed reported for Buckinghamshire upper schools and the revenuefromstatements of SEN to arrive at theadditional cost. subtracted Pastoral care levelsof social deprivation, have a Secondarymodernschools,because of higher need to provide pastoral care services greater fortheir students. Elevenper centof schools. upper school pupils are eligible forFSM comparedto 1.2% in grammar thespecialneedsservice showedthatupperschools, in Buckinghamshire Data from whichenrol56% ofsecondary permanent students, had twiceas many as exclusions as term Quantifying this grammar schools and fivetimes manyfixed exclusions. expense given was problematic in deployment additional theabsenceofdataon staff a conservative estimate schools.Therefore grammar was made thatupperschools schoolson pastoral careand so half thecost haveto spendtwiceas muchas grammar spinepointsforteachers undertaking specific pastoralcare roles of the additional Table 10. Estimates ofadditional school educational costs ofsecondary in provision modem Buckinghamshire incurring cost Area ofactivity additional
courses Vocational Specialneeds and retention Recruitment costs additional KS3 Strategy: costs Additional schools 2002. upper Source:questionnaire to Buckinghamshire
Pastoralcare

Costper pupil
4.15 147.00 13.61 3.84 175.00
5.92

This content downloaded from 132.234.251.230 on Thu, 12 Sep 2013 02:13:39 AM All use subject to JSTOR Terms and Conditions

Aresecondary disadvantaged? 171 modern pupils was included.The revenue from external funding of pastoralcare was subtracted totalcoststo arrive at an estimate from of upperschools'additional costs. Recruitment and retention on how much theyspentper yearon recruitment Schools reported and retention for teachers and on advertising points costs.This amounted to 340,000.Fromother 6% of the teaching responses it was calculatedthatvacancieswereapproximately establishment. Almost12% ofteachers contracts wereon temporary and 11% were qualifications which their inpositions It working orexperience for wereinappropriate. schools is reasonable ingrammar to assume thesefigures wouldbe considerably lower and thata conservative estimate of the additional costsof recruiting and retaining This gavea totaladditional itis twiceas high is that in upperschools. teachers costof 170,000 for recruitment in the 13 upperschools. and retention the school for implementing the budget from Additional KeyStage3 strategy22 contributions which theKey Stage3 Strategy from thecostsofimplementing Schoolsreported the from amountfinanced Fund was subtracted. Standards It was arguedby thehead schoolsrecruited teachers the ableststudents thatas grammar theydid not incur additional costsfortheKey Stage3 Strategy. costofall theactivities The additional to be 175 per listedabovewas estimated upper pupil per from to This compares received 79 pupil. school of SEN without so there formula forthe funding Buckinghamshire's statements, of almost100 per pupil. was an estimated shortfall modern schoolsand of deficit The greater occurrence budgetsamongsecondary lendsupport high to theevidence from itsparticularly incidence in Buckinghamshire on coststhattheseschoolsdo faceadditional thequestionnaire whichthey costsfor werenot adequately funded.
Transport cost incurred by selective systems

Although selective LEAs tend to fundschoolsless generously than non-selective LEAs, theytendto spend moreon home-to-school transport. system A selective costsmoreto operate system to because students thana non-selective are less likely attendthe school closestto their home and therefore requirepublicsubsidiesfor transporting themto school. This is compoundedin ruralareas by the greater An additional whengrammar cost and inequity is perpetrated distancesinvolved. schoolifit is morethanthree schoolpupilsareprovided free with transport to their milesfrom their home,regardless grammar ofwhether ornotitis thenearest school, ifthenearest modern transport whilesecondary pupilsonlygetsubsidised schoolis
over threemiles fromtheirhome. This is the policy withrespectto school transport in Buckinghamshireand it perpetuates an academic pecking order in the county's grammarschools. A regressionof home-to-school transportexpenditureper pupil at LEA level in 2000/01 shows that the proportionof pupils in the LEA attendinggrammarschools

This content downloaded from 132.234.251.230 on Thu, 12 Sep 2013 02:13:39 AM All use subject to JSTOR Terms and Conditions

172 R. Leva`iCandA. J. Marsh to higherexpenditure. Other factorsare low population density contributes (sparsity), needs indicator theLEA's additional the costsof educational (reflecting pupils SEN) and averageschool size (whichis inversely transporting more with related to transport withhome-tocostsperpupil). The association ofthesefactors shown population expenditure has the density per Table pupil is school in 11. Low largesteffect of pupils attending schools.23 by the proportion followed grammar schoolsis associatedwithabout 10 per Every10% of LEA pupilsin grammar annumextraper pupilin transport costs. Conclusions It is clear from the evidence presentedthat secondarymodern schools are to the needs of their underfunded in relation pupilsand thatpupilsare in general intensity ofneed academically by attending them.There is a greater disadvantaged in secondarymodernschools on a numberof measures.The costingexercise provided some indicative evidence from one of the selective LEAs, provision in secondary Buckinghamshire, thatthe additionalcosts of educational modem schoolswerein theregion of175 per pupilin 2000/01.Notwithstanding needsofpupilsin secondary thebudgetshareperpupil modemschools, thegreater schoolswas virtually ofsecondary modemand grammar 2000/ thesamein theyears when is central direct government funding included, 01 to 2002/03.However, the difference had movedin favour in totalpublicrevenue ofsecondary modemschools of the 10 selective to the tune of 100 per pupil by 2002/03.Additional analysis schoolsweremore modern In Sloughand Sutton LEAs showsvariability. secondary generouslyresourced than grammarschools. The reversewas the case in Buckinghamshire, Medway, and Southend,where secondarymodem schools, regardless lessbudget havesixth ofwhether received shareperpupilthan they forms, grammar schools. secondary of whether modernschools are underA more accurateassessment funded is obtained fromregression analysisas this enables us to controlfor
costs:dependent Table 11. Factors variable: associated with transport costper home-to-school pupilperannum 2000/01 Variable Constant in terms (measured ofthe density Low population sparsity theLEA in theSSA formula) for index Proportion ofpupils in grammar schools Average sizeofsecondary schoolin LEA needs(LEA indicator usedin Additional educational SSA formula) coefficient () Estimated 99.65 248.95 103.14 -0.146 16.55

Note: all estimated coefficients aresignificant at .05. 2000/01 Schools 52 Budget statements ofdata:Annual Census2001 andSection Sources Outturn (DfES, 2002b).

This content downloaded from 132.234.251.230 on Thu, 12 Sep 2013 02:13:39 AM All use subject to JSTOR Terms and Conditions

pupils disadvantaged? 173 Aresecondary modem differences in cost drivers.Afteraccountingfor size, sixth form,region cost differences, needs at schooland LEA percentage ofpupilswithadditional learning level,whollyselective LEAs' tendency and Standards to spend less per student, secondary school categories, Fund grantsfordifferent modernschools received around80 less per pupil thancomprehensive schoolsin two out of threeof the schoolsattracted yearsexamined whilegrammar over100 moreperpupilin each ofthethree years. costmorein terms Selective ofhome-to-school that systems transport. Resources could have been allocated to teachingand learningare spent on transporting students to moredistant additional schools.A roughmeasureis thatevery 10% of pupils attending with associated is schools in an LEA grammar an increasein student transport of10 per for expenditure in theauthority. One method reducing transport or at least stabilising the home-to-school budgetwould be to require to schoolsnotin thereserved schoolpupilsto payfortransport parents ofgrammar area,as is the case withparents modern pupils. ofsecondary modern underfunding in the Secondary schools' comparative was also reflected theyhave experienced in keepingtheirbudgetsbalanced. A higher difficulties ofsecondary compared with proportion to comprehensive schoolsoperated modern budgetsin the threeyears2000/01-2002/03. deficit In contrast, a much lower ofgrammar difficulties. proportion schoolswas in financial value-added shows students modem schoolsare analysis The that in secondary oftheir disadvantaged ofattending in terms GCSE/GNVQ attainment. The penalty a secondary modemschoolis on average1 gradeless thanthestudent wouldhave ofgetting 5 obtained at a comprehensive schooland abouta 3% reducedprobability or moreA*-C GCSE/GNVQ grades. Whenwe takeaccountoftheadditional social modem students assignedto secondary systems, segregation createdby selective disadvantaged school to thoseadvantaged schoolsare further relative by grammar due to peergroupeffects. selection byabout1.8 grades The mostablepupilsareless bygrammar advantaged ofadditional in terms GCSE attainment schoolattendance with studies by Schagen and consistent than those of averageability-finding Schagen(2003) and Bentonetal. (2003). in thisarticle that The evidence presented showsclearly system in 10theselective works to disadvantage 20 local authorities pupilswho end up in secondary modem against the resourcing policiesactuallydiscriminate schools. In a fewauthorities secondary the to grammar schoolpupils.This is modempupilscompared downside discretion education. oflocal political to chooseselective system is strong From an equity pointof viewthe case againsta selective but it popularwithparents access appearsunderstandably whosechildren gainprivileged to grammar schools.If the selective system is to remainbecause of local political
support, then there is a clear case for ensuringthat secondary modern schools are betterresourced than grammarschools so as to offset the compounding effectsof lower income and lower abilitystudentsin schools reservedforthem. concentrating The introduction by the DfES of dedicated schools budgets in 2006/724makes such even less likelyto occur throughLEA action thanbefore,since LEAs a redistribution

This content downloaded from 132.234.251.230 on Thu, 12 Sep 2013 02:13:39 AM All use subject to JSTOR Terms and Conditions

174 R. LevacicandA. J. Marsh are required to all to pass on DfES determined in funding realper pupil increases cannot reallocate fundingfromgrammarschools to schools. They therefore raising their on education, own spending secondary which modern schoolswithout to do once education funding In theyare unlikely is no longer responsibility. their expensive the the longer term it is probablyless to raise and more effective to secondary achievement of pupilswho would be consigned modem schoolsby performance givencurrent selection, though dismantling thiswould incurcostsin for those no longerprivileged terms of lower attainment by grammarschool provision.

Acknowledgments
We wishto acknowledge theworkof David Newson and Mike Carslawwho were involved research on schoolcosts,and Andrew for Jenkins in the Buckinghamshire assisting withthe statistical analysis.

Notes
1. 2. Middle deemed secondary schools (i.e. thosewiththe highest age ofpupils 14) are excluded. These are schools that classifythemselvesas secondary modem plus any nominally comprehensiveschools in wholly selective LEAs. In the past some schools reporting modem' in theperformance tablesmayhave quite legitimately themselves as 'secondary been recorded as 'comprehensive' or otherwise withinthe Annual Schools Census. In orderto comparisons, the DfES has now included'secondarymodern'as ensurevalid benchmarking one of six options in the new field'intaketype' to be included in the Pupil Level Annual Schools Census (PLASC) (DfES, 2003a). The intake types are Comp-=Comprehensive, SELl=Selective (Grammar), SEL2=Selective (Secondary Modem), SEL3=Selective (Technical), SEL4=Selective (Religion),SPEC= Special. Due to lack of alternative data, Schagen and Schagen (2003) used Key Stage 2 in 1997 as for Key Stage 3 test resultsin 2000, and Key Stage 3 in 1998 as prior prior attainment forGCSE results in 2000. The problemwiththesemeasuresofvalue added is that attainment highervalue added at Key Stage 4 may be due to catchingup on lowervalue added at Key schoolperformance measureofsecondary is value added from KS2 to Stage 3. Hence a better KS4. We use thismeasurein our analysis. at Key Stage 2 for measure of prior attainment Benton et al. (2003) used the preferable assessingvalue added at Key Stage 3 and at GCSE in 2002, as well as usingpriorattainment at Key Stage 3 2000 linkedto GCSE in 2002. variablemodel was fitted in orderto correct The resultswere weakerwhen an instrumental forbias due to livingin a selectivearea being relatedto family background. least squares regression is biased downwards The standarderror in an ordinary usingnested ordereddata such as pupils in schools). Hence estimatesmay be data (i.e. hierarchically significant reported as statistically when theyare not. The data set used is takenfrom398,532 matchedrecordsfroma representative sample of pupils at Key Stage 2 (1996) to GCSE/GNVQ 2001 supplied by the QCA. Independent school pupils, pupils in special schools and pupils withmissingdata or whose schools had missingdata were dropped,leaving334,574 cases. All these are reputablenationaldata sets widelyused in UK educationalresearch.

3.

4.

5. 6.

7.

8.

This content downloaded from 132.234.251.230 on Thu, 12 Sep 2013 02:13:39 AM All use subject to JSTOR Terms and Conditions

pupils 175 Aresecondary modern disadvantaged?


9. 10. 11. Key Stage 2 attainment was enteredwithlinear,squared and cubed termsas all threeare statistically significant. forlaterintakeyearsthanYear 7 weredroppedas they werenot statistically Dummyvariables significant. betweenKS mathematics and theDurham PIPS mathstestin anyone yearis The correlation 0.85 and for English 0.83 (Tymms, 2004, p. 480). Stobart (2001) notes that the 'test development but thatthe validity of the nationaltestsas measuresof processis exemplary' forour yearto yearchangesin standardsis questionable.However,thisissue is not relevant analysiswhichuses data fora singlepupil cohort. In the regressionsthe continuous variables were standardised but in the tables the coefficients have been convertedto theirnaturalunits (e.g. GCSE total points) to make easier to understand.Hence a coefficient of 0.3 on the Key Stage 2 theirinterpretation average score means that one extra mark at Key Stage 2 predicts 0.3 extra points at GCSE. As the regression equation was estimatedin standardised on the Key units the coefficient termis -4.94 forpupilswhose Key Stage 2 scoresare 1 Stage 2-grammar school interaction standarddeviationabove the mean. These pupils would also lose a further estimated1.04 GCSE pointsby attending modernschool,makingtheirtotalGCSE points2.2 a secondary less than iftheyhad attendeda comprehensive school. Because theseestimates regression theyare non-linear are derivedfrom withrespect a logistic to the explanatory ofa pupilwitha givenKey Stage variables.This meansthattheprobability 5 or more A*-C grade GCSEs varies also with the values of all the other 2 score getting independentvariables,so we need to assume values forthese when presenting illustrative results.In theexamplesgivenin Table 5 we comparepupilswho attendnon-denominational, non-specialist schools,withsixthforms and whichhave averagevalues forall the continuous We thentakea pupil ofgivenability independent variables. that and comparetheprobability and secondary he/she obtains5 or moreA*-C GCSEs in grammar modernschoolscompared to comprehensive schools. The school budgetshare consistedof revenuesdelegatedby the LEA plus standardsgrant and standards fundswhichcome via theDfES. Data on pupilsweretaken from theTreasury from the Annual Schools Census. The pupil rollused to calculatethe budgetshareper pupil 2000 roll plus seven-twelfths of the January for2000/2001 was five-twelfths of the January 2001 roll. The Area Cost Adjustment Factor is used by the Government in calculating LEAs' Education Standard Spending Assessment(Education Formula Share since 2003/ and variesfrom1 to in regionalgrossaverageearnings 04). It is an indexbased on differences 1.25. in regionalcostscurrent revenueper pupilwas dividedbythe To take accountof differences Area Cost Adjustmentfactor since there are too few values of ACA to use it as an variable. independent are givenin the Register of Educational Establishments. The different categories incomepupilsare likely to generate These are publicrevenues.Grammar schoolswithhigher more own revenueper pupil than secondarymodern schools as in Buckinghamshire (see Levadic et al., 2002). Chi-square tests Value Pearson chi-square n of valid cases 49.68 2851 df 6 Asymptotic. sig. (2-sided) .000

12.

13.

14.

15.

16.

17.

18. 19.

20.

This content downloaded from 132.234.251.230 on Thu, 12 Sep 2013 02:13:39 AM All use subject to JSTOR Terms and Conditions

176 R. Levac'id andA. J. Marsh


schoolsas it was anticipated theymight was not administered to grammar 21. The questionnaire had a directcomparison be reluctant to respond.The evidencewould have been stronger schools been possible. withgrammar to improve 22. The Key Stage 3 Strategy was a DfES initiative, teaching and backed by funding, fromentry to secondaryschool to the KS3 tests learningand therefore studentattainment threeyearslater. 23. This can be seen from regression the sizes of the standardised coefficients. 24. In December 2004 the DfES announced thatLEAs must pass on a mandatedpercentage increaseper pupil to all schoolsand thatfrom 2006/07LEAs willbe givena dedicatedschools grantto pass onto schools.

References
performance ofmaintained Benton,T., Hutchinson,D., Schagen,I. & Scott,E. (2003) Studyofthe secondary schools inEngland.ReportforNationalAuditOffice (Slough,National Foundation forEducational Research). Crook, D., Power, S. & Whitty, a review on of research G. (1999) The grammar schoolquestion: (London, Institute ofLondon). comprehensive and selective education ofEducation,University (London, DfES). DepartmentforEducation and Skills(DfES) (2001 a) Schools:achieving success DepartmentforEducation and Skills(DfES) (200 1b) Theeducation England). statements, (outturn Regulations guidance(relating 2001 and accompanying to the2000-1financial year) (London, DfES). Department for Education and Skills (DfES) (2002a) Statistics of education: in pupil progress secondary byschool schools type in England:2001. National Statistics Bulletin,Issue 05/02. (England). (outturn forEducation and Skills(DfES) (2002b) Theeducation Department statements) year) (London, to the2001-2 financial 2002 and accompanying guidance(relating Regulations DfES). DepartmentforEducation and Skills (DfES) (2003a) Pupil levelannual schools census(PLASC) notes Availableonline at: www.dfes.gove.uk/ forsecondary January2003. Completion schools. datacollection/asc/ASC2003Home.asp. (outturn England). DepartmentforEducation and Skills(DfES) (2003b) Theeducation statements, year) (London, guidance(relating Regulations 2002 and accompanying to the2002-3 financial DfES). Departmentof Education and Science (1965) The organisation Circular/65 ofsecondary education. (London, HMSO). Dustmann, C., Rajah, N. & Van Soest, A. (2003) Class size, educationand wages, TheEconomic 113(485), 99-120. Journal, Edwards, T. & Tomlinson, S. (Eds) (2002) Selectionisn't working. Diversity, standards and Forum). inequality in secondary The Available online at: Catalyst education(London, www.catalystformum.org.uk. Feinstein,L. & Symons,J. (1999) Attainment Economic Papers,51, in secondaryschool, Oxford 300-321. Galindo-Rueda, F. & Vignoles,A. (2004) The heterogeneouseffect of selection in secondary Institute the changingrole of ability, schools: understanding forthe Studyof Labor (IZA), Discussion paper 1245, Bonn, Germany. oftheselective system education ofsecondary Gallagher,T. & Smith,A. (2000) Theeffects in Northern Ireland). Northern Ireland.Main Report (Bangor,Education Department models Goldstein,H. (1995) Multilevel statistical (London, Arnold). diversity House of Commons Education and Skills Committee(2003) Secondary of education: provision. HC 94 (London, HMSO). ofLEAs and schools of differing Jesson,D. (1999) Evaluating (York, Centre for performance types Evaluationand ResourceManagement,Department ofEconomics,University Performance of York).

This content downloaded from 132.234.251.230 on Thu, 12 Sep 2013 02:13:39 AM All use subject to JSTOR Terms and Conditions

Aresecondary disadvantaged? 177 modern pupils


Jesson,D. (2000) Furtherevidence on comparative betweenselectiveand GCSE performance schools and LEAs, paper presentedat NUT Secondary non-selective EducationConference, Harrogate,1 March. Jesson,D. (2001) Selective systemsof education-blueprint for lower standards?Discussion paper. Available online at: http://www.york.ac.uk/depts/econ/rc/cperm.htm (accessed 14 September2001). multilevel modelling (London, Sage). Kreft, I. & De Leeuw, J. (1998) Introducing costsofupper school funding: Levaiic, R., Marsh,A. & Newson, D. (2002) The penalty towards fairness in the secondarysector.Availableonline at: htpp://www.missingbucks.org. greater data report Officefor Standards in Education (Ofsted) (2002) National summary for secondary 1.2 (London, Ofsted). Available online at: www.ofsted.gov.uk/ schools 2001. Data version public/docs01. skills Organisation forthe forEconomic Cooperationand Development(OECD) (2003) Literacy oftomorrow-further from PISA 2000 (Paris,OECD/UNESCO). Availableonline world results at: www.pisa.oecd.org. Organisation for Economic Cooperation and Development (OECD) (2004) Learningfor from PISA 2003 (Paris, OECD). Available online at: world:first results tomorrow's www.pisa.oecd.org. ofnationalvalue added data setsto assess theimpactof Schagen,I. & Schagen,S. (2003) Analysis 29(4), 561-582. Educational Research Journal, selectionon pupil performance, British Shuttleworth, ofperformance at GCSE. ResearchPaper SEL3.1 I. & Daly, P. (2000) Thepattern Ireland). (Bangor, Departmentof Education,Northern and thesocialorder Simon, B. (1991) Education 1940-1990 (London, Lawrence & Wishart). Commission(2005) Measuring schools. primary standards Statistics inEnglish Commission Statistics ReportNo. 23, 10. Availableonline at: http://www.statscom.org.uk/C_402.aspx/. schools (London, National Children'sBureau). in secondary Steedman,H. (1980) Progress Steedman,H. (1983) Examination schools: findings results and non-selective inselective fortheNational Study(London, National Children'sBureau). ChildDevelopment British ofEducational Stobart,G. (2001) The validity of nationalcurriculum Journal assessment, Studies, 49(1), 26-39. Research schools?British Tymms, P. (2004) Are standardsrisingin Englishprimary Educational 30(4), 477-494. Journal,

Appendix

Table Al. Descriptivestatistics of the main continuousvariables Mean K2 total marksforEnglish,mathsand science GCSE totalpointsscore 2001 FTE pupils: average1997-2001 FSM: averagepercentage1997-2001 SEN withstatements: averagepercentage1997-2001 White: averagepercentage1997-2001 PTR average 1997-2001 Number of pupils in school in QCA matcheddata set 1996-2001 (excludes middleand special schools) independent, 154.5 40.4 1056 16.4 2.5 88.4 16.9 340124

This content downloaded from 132.234.251.230 on Thu, 12 Sep 2013 02:13:39 AM All use subject to JSTOR Terms and Conditions

andA. J. Marsh 178 R. Levac'zi

Table A2. Estimatedregression equations forGCSE/GNVQ total score and probability of 5 or more A*-C GCSE grades Independentvariables GCSE/GNVQ total score: (no interactions) naturalunits of 5+A*GCSE/GNVQ Probability total score: (with C (logitregression) interactions) normalisedunits naturalunits

Estimate t statisticEstimate t statistic Estimate t statistic Constant KS2 total forEnglish,mathsand science Age Girl Percentageof girlpupils: 2000 FTE pupils: average 1997-2001 FSM average% 1997-2001 FSM average (1997-2001) squared FSM average (1997-2001) cubed KS2 Englishmathsand science squared average% SEN withstatements: 1997-2001 White: average% 1997-2001 PTR average 1997-2001 Highestage 16 (= 1) Specialistschool Church of England Roman Catholic Otherreligion Grammar Secondarymodern Grammar*KS2total Secondarymodern*KS2 total WhollyselectiveLEA -3.34 0.323 0.009 3.375 0.014 .0002 -0.146 -19.22 323 52 87.5 4.01 0.8 -29.2 -3.32 0.328 0.009 3.375 0.015 .000 -0.146 -19.11 328 51 87.5 4.33 0.8 -29.2 -0.799 2.066 0.135 0.519 0.046 0.001 -0.589 0.174 -0.024 -0.024 -0.038 -0.204 -0.021 0.155 0.131 0.039 0.088 0.568 1.252 -0.164 -28.32 158.92 27.00 47.18 4.18 0.07 -28.05 9.16 -4.80 -3.00 -2.71 -14.57 -1.62 5.17 4.23 0.74 2.51 2.87 14.73 -2.78

0.018 -0.20 -0.08 -0.27 0.89 1.47 0.89 0.96 2.08 5.50 -1.04

36 -4.25 -16.2 -3.75 4.6 7.6 2.56 4.17 1.96 10.6 -2.16

0.019 -0.017 -0.08 -0.27 0.89 1.48 0.89 0.94 2.51 9.37 -1.23 -4.94 -1.04 0.02

39 -4.25 16.2 -3.75 4.6 7.70 2.56 4.08 2.36 16.76 -2.56 -18.29 -6.75 0.03

on KS2 attainment Note. The coefficient is random. In Table A2 we presentthe estimatedcoefficients fromthe logitequation in standardised units, of the dependentvariable (5 or more A*-C GCSE/ whereasin Table 5 we give the probability GNVQ) forgivenvalues of the explanatory of a pupil getting variables.The probability 5 or more forgiven A*-C GCSE/GNVQ varieswiththe values of the explanatory variablesand is reported assumptionsabout theirvalues. Given thatp standsforthe probability thata pupil obtains5 or equation is the natural more A*-C GCSE/GNVQ, the dependentvariablein the logitregression x is theexplanatory variableand b logarithm ofp/(1-p),i.e. In[p/(1-p)]=ebxwheree is exponential, on x. The probability, the coefficient p, of 5 or more A*-C GCSE/GNVQ is then calculatedby working out 1/[l+e-bx].

This content downloaded from 132.234.251.230 on Thu, 12 Sep 2013 02:13:39 AM All use subject to JSTOR Terms and Conditions

You might also like