selected papers in school finance 1995 · 2 selected papers in school finance, 1995 introduction...

NATIONAL CENTER FOR EDUCATION STATISTICS

Selected

Papers

in

School

Finance

1995

U.S. Department of EducationOffice of Educational Research and Improvement NCES 97-536

U.S. Department of EducationRichard W. RileySecretary

Office of Educational Research and ImprovementMarshall S. SmithActing Assistant Secretary

National Center for Education StatisticsPascal D. Forgione, Jr.Commissioner

The National Center for Education Statistics (NCES) is the primary federal entity for collecting,analyzing, and reporting data related to education in the United States and other nations. It fulfills acongressional mandate to collect, collate, analyze, and report full and complete statistics on thecondition of education in the United States; conduct and publish reports and specialized analyses ofthe meaning and significance of such statistics; assist state and local education agencies in improvingtheir statistical systems; and review and report on education activities in foreign countries.

NCES activities are designed to address high priority education data needs; provide consistent,reliable, complete, and accurate indicators of education status and trends; and report timely, useful,and high quality data to the U.S. Department of Education, the Congress, the states, other educationpolicymakers, practitioners, data users, and the general public.

We strive to make our products available in a variety of formats and in language that is appropriate toa variety of audiences. You, as our customer, are the best judge of our success in communicatinginformation effectively. If you have any comments or suggestions about this or any other NCESproduct or report, we would like to hear from you. Please direct your comments to:

National Center for Education StatisticsOffice of Educational Research and ImprovementU.S. Department of Education555 New Jersey Avenue, NWWashington, DC 20208-5574

Suggested CitationU.S. Department of Education. National Center for Education Statistics. Selected Papers in SchoolFinance, 1995, NCES 97-536, by Fowler, William J. Jr., Washington, DC: 1997.

You are welcome to reproduce any part or all of this publication since the government holds nocopyright on the materials it publishes. However, if you use one of our papers, please credit theNational Center for Education Statistics (NCES).

March, 1997

Contact:William J. Fowler, Jr.(202) 219-1921Internet:[email protected]

This publication is dedicated to Charles S. Benson, a distinguished economist andeducator at the University of California, Berkeley, who dedicated his life’s work to improvingeducation finance for students in poor and downtrodden school districts, which he believedtrained them to wander from one minimum wage job to another.

Dr. Benson was a professor emeritus whose three decades at Berkeley were spent teach-ing educational administration and policy analysis. He also was the director of the NationalCenter for Vocational Education, which studied the effects of education on employment. Histext, “The Economics of Public Education,” is still widely acknowledged as a classic in thefield. Dr. Benson, a native of Atlanta, Georgia, received his bachelor’s degree in economicsfrom Princeton University and his master’s and Ph.D. from Columbia University. Beforejoining the Berkeley faculty, he taught at Bowdoin College from 1950 to 1955 and at HarvardUniversity from 1955 to 1963.

Although the editor did not have the pleasure of knowing Dr. Benson personally, thosewho did remember an erudite, trenchant, and amusing fellow whom others sought out asmuch for his humor as his insight and wisdom. The editor wishes to express his appreciationof Dr. Benson’s lifelong efforts to change existing systems of school finance in distressedschool districts. As G. Alan Hickrod confides in this volume, Dr. Benson testified before aU.S. Senate Committee:

You must be very careful when you wish for things because you may just getwhat you wish for. We worked hard for equity in California. We got it. Now

we don’t like it.

Those contemplating a life journey similar to that of Dr. Benson, striving to improve thefinancing for education of students in distressed school districts, would do well to remembernot only his enormous contribution to the education finance community and to students, butalso his style, wit, wisdom, and especially, his kindness and willingness to help all those heencountered.

Dedication

In memory of Charles Scott Benson,

1922–1994

i

ForewordPaul D. Planchon, Associate Commissioner

Elementary and Secondary Education Division

The National Center for EducationStatistics (NCES) constantly reevaluates itsefforts in the field of school finance bycommissioning papers from distinguishedmembers of the school finance researchcommunity, asking them to assess the dataneeds of the profession. Even when thesedata needs have been satisfied, a number ofdifficult statistical and measurement ques-tions arise when conducting empirical andquantitative research. The papers presentedhere were commissioned by NCES toaddress the twin concerns of what additionalschool finance information NCES shouldcollect and report, and how extant data

might be analyzed to address interestingquestions faced by the profession.

This report is the second in the renewalof this series, which previously was discon-tinued in 1977. The papers are intended topromote the exchange of ideas amongresearchers and policymakers. Because theviews are those of the authors, the papersmay provoke discussions, replications,replies and refutations. If so, the publicationwill have accomplished its task. Therewould be nothing so satisfying to the Centeras promoting and contributing to the field ofschool finance.

iii

AcknowledgmentsThe editor wishes to gratefully acknowledge the comments and suggestions of the

reviewers: Lee M. Hoffman and Michael P. Cohen of the National Center for EducationStatistics (NCES). I also wish to acknowledge the contributions of Mia Perona, RebeccaPratt, and Phylis Greenfield of Pinkerton Computer Consultants, Inc. who edited the manu-script and incorporated the text and graphics into a published document.

v

Table of ContentsDedication

Foreword .............................................................................................................. iPaul D. Planchon

Acknowledgments .......................................................................................... iii

Introduction and Overview ............................................................................ 1William J. Fowler Jr.

Does Money Matter in Education? A Policymaker’s Guide ................. 15Lawrence O. Picus

The Effect of Constitutional Litigation on EducationalFinance: A Further Analysis ........................................................................ 37

G. Alan Hickrod

Student-Level School Resource Measures ................................................. 61Robert Berne and Leanna Stiefel

Proposed “Good Practices” for Creating Data Bases fromthe F-33 and CCD for School Finance Analyses ....................................... 83

Michael O’Leary and Jay Moskowitz

The Empirical Argument for Educational Adequacy, theCritical Gaps in the Knowledge Base, and a SuggestedResearch Agenda........................................................................................... 101

William H. Clune

Introduction and Overview

Selected

Papers in

School

Finance

2 Selected Papers in School Finance, 1995

Introduction and Overview

Dr. William J. Fowler, Jr. is an educa-tion statistician at the U.S. Department ofEducation’s National Center for EducationStatistics (NCES) who specializes in schoolfinance and educational productivityresearch. His work has centered on rede-signing the federal school finance datacollection to obtain information that canprovide more policy-oriented analyses forthe school finance community. NCES hasreinstituted a state and school districtfinance data collection for the first time inmore than a decade, and is currentlyfunding exploratory research work.

Prior to his work at NCES, Dr. Fowlerserved as a supervisor of school financeresearch for the New Jersey Department ofEducation and taught at both Bucknell

University and the University of Illinois. Healso served as a senior research associate forthe Central Educational Midwestern Re-gional Educational Laboratory (CEMREL) inChicago and for the New York Department ofEducation.

Dr. Fowler has been a member of theAmerican Education Finance Associationsince 1977, and was elected to its Board ofDirectors in 1992. He is a coauthor ofDisparities in Public School Spending, 1989–90, and a coeditor of Organizational Influ-ences on Educational Productivity, publishedby the JAI Press, and serves on the editorialboard of the Journal of Education Finance.He obtained his doctorate in education fromColumbia University in 1977.

About the Author

Introduction and Overview 3

The National Center for EducationStatistics (NCES) commissioned the papers inthis publication to address certain perplexingquestions in education finance. Earlier papersin this NCES series focused on the nation’seducation finance information needs andconcerned emerging education finance topicsthat posed statistical and measurementproblems for the profession. This publicationextends that tradition by first examining twopolicy related issues, whether money mattersin education and the effect of state constitu-tional litigation. Certainly those involvedwith education policy have struggled tounderstand how money matters in educationand must be astonished by educationalresearch that finds no strong or consistentrelationship between the two. Similarly, withmore than 40 states having had the constitu-tionality of their state education fundingsystems challenged, and with a plaintiffsuccess rate of about 50 percent, thoseinvolved with education policy must wonderif it is worth challenging a state fundingsystem through the courts.

Additional papers explore three statisticaland measurement problems that NCES hasencountered. The first is how to measureresources at the student level rather than at theschool or school district level, which is the

custom today. NCES data bases on asample of the nation’s students containimpressive information about the educa-tion of the student, with the exception ofthe resources devoted to that student.Another statistical and measurementproblem is what constitutes “goodpractice” when conducting educationfinance research using NCES data bases.Researchers tend to be idiosyncratic intheir approach to preparing educationfinance data bases for analysis, but mighta common approach be appropriate? Thethird, and most ambitious paper, struggleswith a suggested research agenda formeasuring educational adequacy. Toobtain these papers, NCES turned todistinguished education finance research-ers. NCES asked that they turn theirknowledge, experience, and insightstoward examining these issues, and to puttheir thoughts in publishable form.

Perhaps no more perplexing questionarises than that of whether money mattersin education. The first paper in thispublication, by Lawrence O. Picus ofthe University of Southern California,explores the troubling finding that “thereis no strong or systematic relationshipbetween school expenditures and student

Introduction andOverview

William J. Fowler, Jr.National Center for Education Statistics


performance” (Hanushek, 1989) and the morerecent work that has questioned that conclu-sion (Hedges, Laine and Greenwald, 1994).Rather than focusing on the complex statisticalanalyses and econometric production-functionapproaches, Picus reviews existing studies thatattempt to ask the question: “Does MoneyMatter?” and considers alternative approachesto the question of whether or not resourcesmight enhance achievement. First, Picusbriefly presents the overall pattern of expendi-tures for elementary and secondary educationduring the last century, along with evidence onhow student achievement has changed overtime. He then turns to the debate betweenHanushek and Hedges, et al. Finally, he offerssome conclusions and policy recommenda-tions.

Few policy analysts realize that, accordingto Hanushek, expenditures on education havegrown faster than spending for health care,and represent some 3.6 percent of the GrossNational Product (GNP) in 1990 (compared to1 percent a century ago). This money hasgone predominantly to hire more teachers, payhigher teacher salaries, and lower class sizes.Although it is often thought that the class sizesare the product of educating children withdisabilities, Hanushek and Rivkin suggest onlyone-third of the recent decline in class sizes isaccounted for by special education programs.Where has the other money gone? Hanusheksuggests it has been used to pay teachersincreased benefits, such as teacher retirement,unemployment compensation, social securitycontributions, and group insurance, such ashealth insurance.

As Picus points out, school districts spendan average of 60 percent of their current fundson instruction, and they show little variation inthat percent. As many have observed, thesubstantial increase in education funds citedby Hanushek has not been matched by higherstudent outcomes on the average ScholasticAssessment Test (SAT) or on the NationalAssessment of Educational Progress (NAEP).

The evidence for graduates to find andkeep good, high paying jobs, however, ismore encouraging. Card and Kruger(1990) found that students with smallclasses and higher teacher salaries hadhigher earnings in their adult years.

Picus begins his review of production-function studies of whether or not moneymatters in improving education by pointingout that students’ socioeconomic status(such as their parents’ education, occupa-tion, and income) may be more importantin determining how well they do in schoolthan are many of the kinds of inputsconsidered in these econometric studies.Typical inputs that are examined for theirrelationship to student achievementinclude per-pupil expenditures, pupil/teacher ratios, teacher education, experi-ence and salary, school facilities, andadministrative arrangements.

The debate between Hanushek andHedges, et al., is whether or not Hanushekshould have included all the studies whichhad insignificant results, or for which thedirection of the effect could not be deter-mined. Even when Hedges, et al., excludethe studies which they believe to beinappropriate, they conclude that whilemoney might matter, class size and teachercharacteristics may not. Hanushek thenasks, if this is their finding, what factors doin fact matter, since teachers’ salaries andclass size are the two largest determinantsof spending. In a recent answer toHanushek’s question, Ferguson (1991)conducted research in Texas that leads himto conclude that hiring more and betterteachers does lead to higher student testscores.

Picus concludes by observing thatwhat we don’t know is what the impact onstudent performance would be if schools orschool districts were to dramaticallychange the way they spend the resources


available to them. Undoubtedly, this debatewill continue, although as Picus suggests, itmay be that the wrong question is beingasked. Perhaps instead the question shouldbe “How is Money Used in Education?”

The second paper, by G. AlanHickrod, Distinguished Professor Emeritusat Illinois State University, examines theeffect of constitutional litigation on educa-tional finance. Some 44 states have experi-enced such litigation challenging their stateaid to education systems. Hickrod reviewssix empirical studies of either the effect ofconstitutional challenges, or studies thatprovide some data for the investigation ofthat topic. Hickrod (1992) studied the1970–90 time period, and concluded thatwinning litigation in a state supreme courtsaw only a modest increase in per-pupilexpenditures, but shifted the tax burdenfrom state to local resources. Winning didseem to reduce disparities between schooldistricts in spending per pupil. However,Heise (1995) did not find an increase in perpupil expenditures following the plaintiffs’successes in either Wyoming or Connecti-cut. Peternick (1995) studied 12 stateswhere the state supreme court had found thestate education aid system unconstitutional,and concluded that these decisions didstimulate growth in per-pupil expenditure.These studies differ in that Heise/Peternickassume some single, sharp increase, whileHickrod believes that the state legislature isfaced with a winning court decision,compliance litigation, and increasingpressure to modify the state aid system in away that is beneficial for the successfulplaintiff school districts.

What makes the research so difficult isthat plaintiff challenges might be character-ized in different ways, such as “a clearwin,” or “a loss,” followed by yet anotherfiling, under yet another legal theory. Inother states, litigation has been filed, but nohearings have been held or decisions

rendered. Assigning “climate” to this bodyof litigation is not precise. Hickrod examineswhether or not there might be the effect thatif a state chooses to reduce disparity inspending between school districts (an equitygoal), that perhaps the overall funding levelof the state will suffer (an adequacy goal).He finds no significant relationship betweenadequacy and equity goals. He also findsthat the data support the notion that the statesupreme court decision in favor of theplaintiff school districts reduces expendituredisparity between school districts withinstates, even if they are not related to in-creases in average per-pupil expenditures inthose states.

Hickrod concludes by examining anec-dotal evidence from 11 states. He alsoaddresses some unanswered questions.While Hickrod asserts that constitutionallitigation is one factor in explaining thechanges in state funding for education, it isnot the only factor. However, he confirmsthat it is worth challenging the funding of K–12 education via the courts, albeit that thelegal process can be quite expensive and verytime-consuming. In addition, chances oflosing the case are quite high, as only 17states have successfully defended theirfunding system. Hickrod correctly observesthat public education is like absolutely noother public service, in that it has its ownarticle in every single state constitution inthis nation.

The remainder of the papers explorestatistical and measurement problems thatNCES has confronted. A perplexing di-lemma faced by NCES is how to measurestudent-level resources, perhaps to accom-pany NCES student-level education andpersonal characteristics data, when mostfinancial data reside at the school-districtlevel. Robert Berne is Vice President forAcademic Development at New York Univer-sity, and his coauthor, Leanna Stiefel, isprofessor of economics at the Robert F.Wagner Graduate School of Public Service at


NYU. They are perhaps best known for theirclassic work in The Measurement of Equity inSchool Finance (1984).

The purpose of the Berne and Stiefelpaper is to discuss the types of student-levelresource measures that would be preferred ifstudent-level resource data were collected on aroutine basis. They begin by exploring thetypes of questions that could be answered withstudent resource data. They then reviewselected literature to show how some educa-tion finance researchers have attempted toanswer such questions with extant data. Theyalso introduce some cost accounting conceptsthat are useful in thinking about how to collectappropriate student-level resource data. Theyconclude by recommending alternatives forNCES to consider.

Berne and Stiefel believe there are threequestions that student resource data couldanswer. The first has been termed productionfunction questions, that is, questions ofresource effectiveness and cost-effectiveness.Examples are: Do additional resources forchildren lead to additional outcomes? andWhat is the cost effectiveness of one programversus another? The second type of questionthat might be answered with these studentresource data are equity questions. Oneexamples is: Whether poor students receivethe same (or greater) resources than otherstudents in a school, or an educational pro-gram, such as preschool education. A thirdtype of question is termed a resource intentquestion. If a student is entitled to specificresources, such as handicapped, bilingual, orcompensatory education, do they receive theadditional resources which were intended tobe assigned to them?

Using illustrative studies, Berne andStiefel demonstrate that “...even when re-searchers are careful to put together a compre-hensive and unique data set, they cannotalways obtain resource variables at the correctunit of analysis.” The reason for this is thatthese studies often use administrative data

because it is available, rather than becausethey prefer such data. In addition, data atthe student level are preferable in produc-tion-function studies to data disaggregatedfrom the school district level. This isparticularly true because student outcomesmight be sensitive to resources not cur-rently available in administrative recordsat the school district level.

Examining equity studies, Berne andStiefel conclude that data problemsimmediately emerge with administrativedata from different sources, mergedtogether for an equity study. For example,pupil counts of students in programs maynot match students funded. Expendituresfor employee benefits, transportation,school lunches, or utilities may be omitted.

Berne and Stiefel include four resourceintent studies in their review. Theseemploy cost accounting methods which areexpensive, and use a bottoms-up approachthat identifies the student and program ofinterest, and then assigns expenditures to achild in that program. The cost accountingmethod is valuable because it emphasizesthe need to conceptualize the use of theexpenditure data before the data arecollected.

Turning to concepts of costing, Berneand Stiefel explain the distinction betweendepartmental and product costingmethods where distinctions are madebetween how full costs can be subdivided,how expenditures can be allocated tostudents, and the difference between“costs” and “expenditures.” Departmen-tal costing finds the costs of administrativeunits, such as school districts, departments,responsibility centers, etc. The primarypurpose of departmental costing is to helpmanagers administer units efficiently.Product costing, in contrast, finds thecosts of producing various kinds of prod-ucts. Berne and Stiefel argue that productcosting is relevant because this is where


most of the questions to be answered withthe resource data are focused.

Berne and Stiefel also argue that thesethree questions require resources linked tostudents, rather than schools or schooldistricts. Products are assigned their fullcosts if the total of all products equals theorganization’s total costs. Direct costs arethose that can be assigned uniquely to oneproduct. Indirect costs, on the other hand,are incurred because of the production ofmany products, and are frequently consid-ered overhead costs. When students are theproduct, instructional supplies are anexample of direct costs, while theprincipal’s time is an indirect cost. Twosystems are generally used for assigningdirect costs to products. Job-order costingdetermines the costs of each individual unitof a product, for example, the cost of fine,hand-made, custom-ordered furniture.Process costing determines the costs ofgroups of identical units and then dividesthe number of units to obtain an averagecost. An example might be the cost of abox of breakfast cereal. Job-order costingis more accurate, but much more difficultand expensive to collect. For education,most likely a mixed model would be used,such as assigning resource costs to types ofstudents, rather than individual job ordercosting.

Berne and Stiefel conclude withrecommendations, framed by the purposefor which the student-level resource mea-sure will be used. They believe the student-level resource measure should be developedto assess the effectiveness of resource use,requiring the linking of resources withstudent outcomes. They argue that NCESshould proceed to develop a student-levelresource measure, regardless of whether thedata are actually collected. They assert astudent-based system (product costing) willbe required. They also urge NCES tocollect full costs, direct versus indirectcosts, at different educational programs.

The measure must include the sources offunding, such as local, state, or federal.Finally, for those costs that must be allo-cated, guidelines must be developed for thebasis and method of allocation. This student-level resource measure, obtained on a sampleof students (like many existing NCESstudent-level data sets), would permit NCESto begin to make progress on the question ofeffective use of resources in education.

The next paper on statistical and mea-surement problems that NCES has contendedwith is by Michael O’Leary of Coopers &Lybrand and Jay Moskowitz of PelavinResearch Institute. They explore some of thesteps that were required to be taken beforeconducting various finance studies fromNCES data sets. Although some of theproblems they cite from the work that wasundertaken in 1994 have been corrected inlater editions of the NCES Common Core ofData (CCD) CD-ROM, the procedures theyrecommend are always appropriate to ensureaccurate results. Although the authorsacknowledge that individual research meth-ods will (and should) always vary, if the databases are constructed with the same sets ofunderlying procedures, researchers’ conclu-sions can be debated on their merits, ratherthan on differences in underlying data sets.O’Leary and Moskowitz begin by pointingout that the school district finance collectionundertaken by the U.S. Bureau of the Censustypically contains a universe of more than16,000 school districts in years ending in 2 or7 (1992, 1997). In other years, the data arecollected only for school districts larger than15,000 students, or if the school district isfiscally dependent upon a county or city.

The number of school districts notprocessed by Census also varies by the yearof the sample. NCES also requested thatCensus collect and process a universe ofschool districts in 1990, to accompanyanother NCES project, mapping decennialcensus demographic information (such asmedian income) to school district boundaries.


O’Leary and Moskowitz also explain thatthe NCES Common Core of Data (CCD) CD-ROM contains data at three different levels(state, school district, and school) from fivedifferent NCES surveys. The CCD CD-ROMtypically contains five years of data for eachof these levels. As a result of this complexstructure, O’Leary and Moskowitz find threeissues arise: sampling issues; school districttypes, and school district levels. Educationfinance researchers have sometimes spentconsiderable time and resources exploringwhy revenue data did not exist for manyschool districts over a decade. Other research-ers weighted non-missing districts, or imputedvalues for missing districts, although the effecton the research results is unknown. There arealso a variety of “special” school districts,including community colleges, vocational-technical school districts, correctional/custo-dial school districts, special education, andnon-operating school districts. Non-operatingdistricts have students but no buildings, andtypically transfer their resident children toother school districts to be educated. Thesedistricts are typically not included in equityand education finance analyses. O’Leary andMoskowitz also explain that there are fourbasic “levels” of districts in the Census F-33collection: elementary only; secondary only;unified K–12, and college-graded. Research-ers are urged not to only deal with a singletype, such as unified, because some states donot have this type of school district organiza-tion, and others have “mixed” models. Differ-ent researchers have approached this differ-ently, although virtually everyone pupil-weights [see the classic work The Measure-ment of Equity in School Finance (Berne andStiefel, 1984).]

O’Leary and Moskowitz also discussunresolved data base creation issues, such asenrollment; special-needs pupils; property,poverty, and income data; and state anddistrict finances. Although NCES and Censusenrollment counts are now likely to be thesame, it is still impossible to separate fundstargeted for special-needs pupils from those

allocated through basic state education aidprograms to school districts. Regardingdistrict finances, NCES has resolved “on-behalf’ funds by imputation for currentyears, and the Governmental AccountingStandards Board (GASB) has included thecollection and reporting of these numbersat the school district level, beginning in1995. However, “outliers,” erroneous andextreme values of per-pupil expendituresor revenues, will always plague datareporters such as NCES, especially whenreporting values for more than 16,000school districts. O’Leary and Moskowitzconclude their discussion with a model of“suggested best practice” when creating ananalysis data set from NCES educationfinance data.

The last paper to explore statistical andmeasurement problems encountered byNCES is a suggested research agenda forexploring educational adequacy, byWilliam H. Clune of the University ofWisconsin-Madison. Clune explains thatthe goal of educational adequacy, definedas high minimum educational outcomes forthe disadvantaged, is constrained by thelack of a strong knowledge base, and that awell-formulated research agenda couldmake an important contribution to educa-tion policy. Clune argues that while betterstudent educational outcomes have neverbeen more highly valued, the link betweenthese outcomes and educational reform andinvestments is widely doubted. This hasthe most poignant effect upon educationfor the disadvantaged, where poor childrencompose the majority of students withpoor educational outcomes. Under suchdoubts, strains are developing in thetraditional goal of equity in educationfinance. Traditional equity argumentsguarantees equitable inputs regardless ofoutcomes. Now cost-effectiveness andproductivity goals intrude. Clune arguesthat if society and the courts becomesatisfied that high minimum educationaloutcomes can be produced at minimum


cost, then this strain will be minimized, andadditional funds will become available.

Adequacy can only be assessed, Cluneargues, when it refers to some outcome.Levin (1991) demonstrated that compensa-tory spending could be based on the ben-efits of good education and the social costsof a poor education. Unfortunately, it isdifficult to be certain that compensatoryspending in the present will reduce crimecosts in the future, particularly for today’staxpayers. Fortunately, Clune asserts,standard setting for education is wellestablished, providing some notion ofeducational adequacy. Turning to theavailable resources to establish adequacy,Clune explains that educational spendinghas grown faster than the cost of living formany decades, and maintains its relativeposition even in the midst of calls forausterity. Very few call for the outrightrepeal of educational subsidies, or therollback of universal high school education.Clune remains optimistic “...that a substan-tial amount of extra money could be foundfor adequacy.”

Clune is skeptical about disagreementin measuring adequacy, although courtsfrequently develop long lists of desirableoutcomes for education (Rose v. Council forBetter Education, Inc., 790 s. w. Ken. 2d186, 212-13 [1989]; and Kentucky andCalifornia student tests result in manystudents unable to attain “proficient” scores(Kirst et al., 1995). He believes that thereis more theoretical than practical confusion,as a simple working definition of adequacyincludes literacy, numeracy and problemsolving, and completion of high school in amanner sufficient to be eligible for furthereducation. Clune then suggests that “highminimum achievement” might be defined asthe attainment of average achievement bydisadvantaged children. This high mini-mum achievement would be equal to theachievement of the average child in thenation. Policymakers, he argues, would

benefit from a fleshing out of the variousoperational definitions of adequacy.

One problem that Clune touches on is theproblem of educational adequacy in high-poverty schools. He asserts that there is stillconfusion regarding whether children thathave not reached educational adequacy may,“within any reasonable range of resources”achieve high minimum performance. Al-though Slavin and his colleagues (Slavin,1994), claim to be able to raise the readingachievement of elementary students toaverage grade levels, no data on costsaccompany this claim. Clune argues that themaintaining of an evaluation data base(including, presumably, costs) for a group of“accelerated schools” would be “ well worththe investment.”

The basic elements of successful acceler-ated education have been known for sometime, Clune asserts, and they are an acceler-ated curriculum, high expectations forstudent learning, a positive school climate,and a safe and orderly environment. Theultimate goal of educational adequacy isaccelerated education, for if poor childrenlearned at the same rate as other children,they would not be educationally disadvan-taged. Thus Clune adds to the evaluationdata base the element of educational process,or teaching technology.

We need a fine-grained understanding ofwhat staff and students do differently in orderto begin understanding the conditions underwhich such desirable behaviors can flourish.Careful investigation of schools adoptingmodels of accelerated education are, there-fore, a promising direction for research.

The New Jersey Supreme Court reasonedthat a good estimate of the cost of an ad-equate education is the resources available tothe state’s most successful students, those inthe highest socioeconomic school districts inthe state. This highlights that the cost of anadequate education is not known. Assume


that the additional costs to educate a child inpoverty are $2,000, the question becomes,$2,000 in addition to what? As is apparent,future education finance research must con-duct the types of studies of accelerated schoolsthat could flesh out the costs for remediating adisadvantaged child. Clune also asserts thatresearch is needed on school aid formulas, inorder to provide sufficient “base” funding andto fully cover the additional costs ofremediation. Clune also believes research isneeded about the type of governance struc-tures capable of tolerating and encouraging thereforms that will bring about educationaladequacy for students in poverty.

Clune concludes by listing the ten researchquestions that need to be addressed by aresearch agenda focused on educationaladequacy. Much of the agenda depends ongathering more data from existing acceleratedschools, which, unfortunately, are not verycommon in the nation. Another strategywould be to turn to useful secondary datasources about educating students in poverty.Not all of this research needs be undertaken bygovernment. Foundations and the voluntarycooperation of research scholars and researchcenters might undertake such challengingeducational research.

Summary

The papers published here are but a singlecomponent of the continuing efforts of NCESto obtain and provide education finance dataof interest and utility to the school financecommunity. The first two papers suggest thatthe debate regarding the efficacy of money foreducation will continue, and that NCES willbe called upon to provide even more financialdata that can inform the discussion. A crucialcomponent of the “Does Money Matter?”argument revolves around “good practice” inconstructing the data base for analysis. Inaddition, the controversy makes the need for a

student-level resource measure even morecritical. The courts, as well as the educa-tion finance community, are likely toattend not only to the efficacy of money,and the effectiveness of state constitutionallitigation, but also to the development of astandard of educational adequacy.

The previous publication, SelectedPapers in School Finance, 1994, assertedthat NCES also wishes to make knownconceptual and methodological advancesin the field of education finance, forresearchers and students alike to emulate,replicate, disseminate, and enhance. Thatprevious volume described the evolution ofthis series of papers, and other NCESproducts, such as the CCD CD-ROM,which have provided education financedata in a readily accessible form for allschool districts in the nation. At that timeNCES had also just started work on ageographic cost adjustment (Chambers andFowler, 1995), and will soon release aneven more sophisticated geographic andinflation adjustment (Chambers andFowler, forthcoming).

New Developments

The education finance reporting needsof Congress, the Department, states, andthe education finance research communityare constantly shifting and expanding, asare financial reporting standards, renderingNCES surveys and reports outdated,although they were previously thought tobe satisfactory for at least a future decade.This turbulent environmental press hascreated the need for NCES to have thecapability to proactively undertake “devel-opmental analysis.”

The Education Finance StatisticsCenter (EFSC) is a specifically-designedEducation Statistical Services Institute(ESSI) component to carry out the NCES


need for developmental analysis to increaseNCES’ knowledge, capacity, data collec-tion, and reporting in education finance.Concomitant with developmental analysis,the EFSC might share advances with statesfacing similar finance dilemmas, offeringtechnical assistance to improve their fiscalreporting systems.

The EFSC is currently conducting threedevelopmental activities: assessing howaccounting changes effect NCES surveysand financial reporting; devising experi-mental measures in education finance, suchas measuring the cost of educating studentswith special needs; and assessing the needfor technical assistance to states. A smallportion of its budget is to administrativelycarry out these three developmental activi-ties.

The activities started by the EFSC,while addressing some of the immediatedefinitional, measurement, collection andreporting dilemmas and policy choicescurrently faced by NCES, do not capturethe future developmental challenges ineducation finance. This requiresproactively anticipating the next thresholdin education finance: where the professionseems to be headed and what questionsfuture public and policy demands willcreate.

Future EFSC activities need to revolvearound developmental challenges in educa-tion finance, attaining an increase intheoretical knowledge, analytic capacity,data collection frequency/methodology, andfinancial reporting. For NCES to attain thenext threshold in education finance, and toretain preeminence in the field, definitional,measurement, collection and reportingdilemmas will have to be explored throughdevelopmental analysis by the EFSC.

Among the projects NCES anticipatesexploring in the next year are:

1. Anticipating changes in accountingstandards

NCES fiscal surveys will be affected bychanges in accounting standards promulgatedby the Financial Accounting Standards Board(FASB), which changes the IPEDSpostsecondary finance collection and theplanned K–12 private school finance collec-tion. Both the state-level National PublicEducation Financial Survey (NPEFS) and theschool-district-level F-33 collection will beinfluenced when the Governmental Account-ing Standards Board (GASB) acts at thebeginning of 1997.

2. The acquisition, analysis, and report-ing of school-level data

The dilemma is how to conduct nation-wide school-level analyses that will enableNCES to report on administrative overheadand program cost, either by collectingschool-level finance data from state adminis-trative records in those states that have suchinformation or amending the NCES Schoolsand Staffing Survey (SASS) to containresource allocation information.

3. Experimental measures of program-matic expenditures and costs

There is no agreement on how to obtainand measure programmatic expenditures ineducation. Programmatic expenditures arethe expenditures for a program, such ascompensatory education, bilingual education,or handicapped education. Expenditures donot necessarily represent the actual cost of aneducational program—ways to exploredifferences between cost and expenditureneed to be developed.


4. Using geographic and inflationarycost adjustments

NCES is in the process of developing bothgeographic and inflationary cost adjustments.However, how these indexes should be appliedto education finance data requires furtherconceptual development. This project willexamine whether the Digest of EducationStatistics and the Condition of Educationmight use the geographic and inflationary costadjustments in reporting financial data.

NCES has established its own home pageat http://www.ed.gov/NCES and the EFSCalso has a home page on education finance athttp://www.ed.gov/NCES/efsc. From thesesites, individuals can obtain products andpublications, such as CD-ROMs or geographicand inflation cost adjustments; request specificdata sets or download them; review frequentlyasked questions about the NCES educationfinance program; and email NCES staff (seegraphic of EFSC home page).

NCES hopes these developments willfurther assist the education finance communityin its research and data needs.

Does Money Matter in Education?A Policymaker’s Guide

Selected

Papers in

School

Finance


Dr. Lawrence O. Picus is an AssociateProfessor in the School of Education at theUniversity of Southern California (USC).He is the director of the Center for Researchin Education Finance (CREF), a School ofEducation research center whose purpose isto study issues of school finance andproductivity in the wake of the recent schoolreform movement. He is also the presidentof the American Education Finance Asso-ciation.

In his role with CREF, Dr. Picus isinvolved with studies of how educationalresources are allocated and used in schoolsacross the United States. He also hasconducted studies on the impact of incen-tives on school district performance and is amember of a number of professional organi-zations dedicated to improving schooldistrict management.

Dr. Picus is the co-author of SchoolFinance: A Policy Perspective (McGraw-Hill, 1992) with Allan Odden, and ofPrinciples of School Business Administra-tion (ASBO, 1995) with R. Craig Wood,David Thompson, and Don I. Tharpe. Inaddition, he is the editor of the 1995 year-book of the American Education Finance

Association, Where Does the Money Go?Resource Allocation in Elementary andSecondary Schools (Corwin, 1995).

Prior to coming to USC, Dr. Picus spentfour years at the RAND Corporation wherehe worked on a number of research projectsincluding a major study of the effects ofintergovernmental grants on local schooldistrict expenditure patterns. From 1977 to1983, Dr. Picus was a principal investigatorfor the Northwest Regional EducationalLaboratory’s Center for State Policy Studies,and conducted a number of studies on issuesof concern to chief state school officers inthe Northwest and Hawaii. He was also theprincipal investigator for a number of schooldistrict specific studies concerning manage-ment issues such as management informationsystems, pupil transportation, school facilityusage and strategic planning.

Dr. Picus earned a Ph.D. in Public PolicyAnalysis from the RAND Graduate School.In addition, he holds a master’s degree insocial science from the University of Chi-cago, and a bachelor’s degree in economicsfrom Reed College. He has a strong back-ground in research design, statistics andeconometrics, and is an expert in the appli-cation of microcomputers.


Lawrence O. PicusUniversity of Southern California

About the Author

Does Money Matter in Education? 17

There is no strong or systematic rela-tionship between school expenditures andstudent performance. (Hanushek 1989, 47)

Relying on the data most often used todeny that resources are related to achieve-ment, we find that money does matter afterall. (Hedges, Laine, & Greenwald 1994, 13)

Introduction

On the surface, it would seem that theexpenditure of more money for educationwould lead to improved student outcomes.Certainly if one were to ask a school teacheror school administrator, they would say thatmore money would help them provide ahigher quality education, which in turnwould lead to greater student achievement.Few in education would suggest differently,and it is rare indeed for a public schoolofficial to say that they could do more withless money, or even with the same amountof money.

Many important educational programsaimed at improving opportunities for groupsof students with special needs are based onthe assumption that additional resources areessential to their success. Compensatory

education programs such as the Federal TitleI/Chapter I program and funding for specialeducation are the two largest examples ofthese programs. Recently, Clune (1994)suggested that to ensure children in poorschools have the opportunity to achieve athigh levels, it might be necessary to provideas much as $5,000 per student in additionalresources.

Despite this general belief that “moneymatters,” the statistical evidence of a rela-tionship between spending and studentoutcomes has been mixed. During the1980s, Eric Hanushek conducted an in-depthanalysis of education production functionstudies and concluded that there is littleevidence to support the existence of arelationship between the amount of re-sources and student achievement (1981,1986, 1989, 1991). Others, notably Murnane(1991), Ferguson (1991), and most recentlyHedges, Laine, and Greenwald (1994), havequestioned Hanushek’s conclusion, arguingthat money does in fact matter.

To date, this debate has focused oncomplex statistical analyses and detaileddiscussions about whether or not the produc-tion function approach is appropriate for


Lawrence O. PicusUniversity of Southern California

...the statisticalevidence of a

relationshipbetween spendingand student

outcomes hasbeen mixed.


To address these issues, the secondsection of this paper offers a brief discus-sion of the data on which this discussion isbased, tracing the overall pattern of expen-ditures for elementary and secondaryeducation in the United States during thelast century, along with evidence on howstudent achievement has changed over time.The third section digs more deeply into theproduction function analyses on the effectsof resources on student performance,focusing most of its attention on the currentdebate between Hanushek and Hedges,Laine, and Greenwald. This section alsodiscusses a number of alternative ap-proaches that merit consideration in at-tempting to ascertain whether or not there isa relationship between spending and studentperformance. Finally, the fourth sectionoffers some conclusions and policy recom-mendations based on the data presented inthe earlier sections of the paper.

General Trends InExpenditures And StudentOutcome

Expenditure Trends

Hanushek (1994b) shows that realexpenditures (inflation adjusted to 1990dollars) for public K–12 education in theUnited States increased from $2 billion in1890 to nearly $190 billion in 1990. Hefurther points out that growth in expendi-tures for education was more than threetimes as fast as growth in the Gross Na-tional Product (GNP), with the result thatK–12 education now represents some 3.6percent of GNP in 1990 compared to lessthan 1 percent a century before. Hanushekalso states that expenditures on educationhave grown faster than spending for healthcare. This is interesting given the intensedebate over the growth in health care costsand the relatively little attention spendingincreases in education have received.

estimating relationships between spendingand student achievement. Much of thedebate on the effect of resources on studentachievement has been based on the produc-tion function approach. Yet as Monk (1992)points out, most efforts to define an educa-tional production function have failed.Despite the inability to relate educationalinputs to outcomes, the strong belief thatmoney is important to improving schoolperformance maintains a strong following(see for example, Murnane [1991]).

Recently a number of researchers,notably Picus (1994a) and Cooper (1993),have looked closely at how school districtsand school sites use the dollars they actuallyreceive. The most stunning conclusion fromthis work is the consistency in the patternwith which schools spend the funds theyreceive. Across the United States, schoolsspend approximately 60 percent of theirresources on direct student instruction. Thisfigure holds true regardless of how much isspent per pupil, and seems to be consistentacross grade levels. These findings suggestthat the effectiveness of new money onstudent achievement may be limited by thefact that these new resources are used in thesame way as existing resources, limiting thepotential effectiveness of those new dollars.

Given that the United States spent over$250 billion on K–12 public education lastyear, understanding the impact that moneyhas on student outcomes is important. Thepurpose of this paper is twofold:

1. To review existing studies thatattempt to answer the question “does moneymatter?,” and provide an objective analysisof the debate in terms of policy outcomes forpolicymakers; and

2. To consider alternative approachesto answering the question of whether or notadditional resources lead to gains in studentoutcomes.

...growth inexpenditures for

education was

more than threetimes as fast asgrowth in the

GNP, with theresult that K–12education now

represents some3.6 percent ofGNP in 1990

compared to lessthan 1 percent acentury before.


Picus (1994a) showed that real per pupilexpenditures in the United States increasedby nearly 70 percent during the 1960s,almost 22 percent in the 1970s, and over 48percent in the 1980s. The total compoundincrease in educational expenditures be-tween 1959–60 and 1989–90 amounted to206 percent. Table 1 updates Picus’ data,and shows that real per pupil expendituresincreased by over 207 percent between1959–60 and 1991–92. Spending on K–12education represented approximately 2.8percent of GNP in 1960, 4.0 percent in1970, and 3.6 percent in both 1980 and1990.

Table 1 also shows the dramatic varia-tion in spending increases across the 50states and the District of Columbia duringthat 33 year period. At the extremes,expenditures in New Jersey increased byover 410 percent, more than four times asfast as they did in Utah where the increasewas just over 100 percent. Real per pupilexpenditures in New Jersey were only $306higher than they were in Utah in 1959–60.However, by 1991–92 that difference hadgrown to $6,277, and New Jersey spentthree times as much per pupil as did Utah.

What has this additional money bought?The most obvious answer is more teachers.Barro (1992) estimated that teacher salariesaccount for 53 percent of all current spend-ing by school districts. Moreover, heestimated that as districts receive additionalfunds, they spend approximately half onteachers, with 40 percent going to reduc-tions in class size and 10 percent devoted toincreased teacher salaries. To demonstratethe effect of this emphasis on reducing classsize, the Digest of Education Statistics(NCES 1994) shows that nationally, thepupil/teacher ratio in public K–12 schoolshas declined from 26.9 in 1955 to 17.6 in1994. Moreover, the pupil/teacher ratiodeclined every year but one between 1955and 1990, and has hovered between 17.2and 17.6 since 1990.

One reason for this decline is oftenthought to be the increase in the number ofchildren with disabilities. Since thesechildren are more difficult to educate, theyoften are enrolled in much smaller classes.Yet Hanushek and Rivkin (1994) show thatspecial education programs account for lessthan one-third of the recent decline in thepupil/teacher ratio. This means that effortsto reduce regular class sizes have succeededas well in most states despite limited evi-dence that smaller classes substantiallyimprove student learning (see, for example,Glass and Smith [1979]; Hanushek [1986,1989]; and Word et al. [1990]).

Of course if Barro’s estimates arecorrect, then half of the average increase inspending goes to objects other than teachersalaries. One factor that is responsible forconsiderable growth in spending in recentyears has been benefits paid to schoolpersonnel. Hanushek estimates that theseso- called fixed charges grew from 7 percentto 14 percent of total spending between 1960and 1980 (comparable data were not avail-able for 1990). These increases are tied bothto the increased number of teachers andother personnel, and to the growing costs ofproviding benefits such as health care andretirement. There are a number of otherimportant functions that must be consideredin the operation of a school system. Centraladministration, for example, only representssome two to three percent of total expendi-tures, while operations and maintenanceaccount for approximately ten percent ofeducational expenditures. Table 2 provides abreakdown on how the more than 15,000school districts in the United States allocatedtheir funds in 1990–91.

Table 2 shows that school districts spentan average of 60 percent of their currentfunds on instruction. Research by Picus(1993a, 1993b, and 1994a) not only con-firmed that figure, but shows in Table 3 thatthere is very little variation in that 60 percentfigure, despite substantial variations in total

...as districtsreceive additionalfunds, they spend

approximatelyhalf on teachers,with 40 percent,

going toreductions inclass size and 10

percent devotedto increasedteacher salaries.


Table 1.—Changes in real expenditure per pupil in average daily attendance (ADA) in public elementary and secondary schools, by state: 1959–60 to 1991–92 (in constant 1991–92

dollars)

Current Current Expenditures Expenditures Percent Change

Per ADA Per ADA 1959–60 to State 1959–60 1991–92 1991–91 United States $1,765 $5,421 207.14% Alabama

1,134 3,616 218.87 Alaska2,570 8,450 228.79 Arizona1,898 4,381 130.82 Arkansas1,059 4,031 280.64 California1,994 4,746 138.01 Colorado1,863 5,172 177.62 Connecticut2,051 8,017 290.88 Delaware2,144 6,093 184.19 D.C.2,028 9,545 370.66 Florida1,494 5,243 250.94 Georgia1,192 4,375 267.03 Hawaii1,527 5,420 254.94 Idaho1,363 3,556 160.90 Illinois2,062 5,670 174.98 Indiana1,734 5,074 192.62 Iowa1,730 5,096 194.57 Kansas1,636 5,007 206.05 Kentucky1,096 4,719 330.57 Lousiana1,749 4,354 148.94 Maine1,330 5,652 324.96 Maryland1,847 6,679 261.61 Massachusetts1,923 6,408 233.23 Michigan1,952 6,268 221.11 Minnesota2,000 5,409 170.45 Mississippi 969 3,245 234.88 Missouri1,618 4,830 198.52 Montana1,932 5,423 180.69 Nebraska1,585 5,263 232.05 Nevada2,025 4,926 143.26 New Hampshire1,633 5,790 254.56 New Jersey1,823 9,317 411.08 New Mexico1,706 3,765 120.69 New York2,642 8,527 222.75 North Carolina1,116 4,555 308.15 North Dakota1,725 4,441 157.45 Ohio1,717 5,694 231.62 Oklahoma1,465 4,078 178.36 Oregon2,109 5,913 180.37 Pennsylvania1,926 6,613 243.35 Rhode Island1,944 6,546 236.73 South Carolina1,035 4,436 328.60 South Dakota1,631 4,173 155.86 Tennessee1,120 3,692 229.64 Texas1,563 4,632 196.35 Utah1,517 3,040 100.40 Vermont1,618 6,944 329.17 Virginia 1,2904,880 278.29 Washington 1,9785,271 166.48 West Virginia 1,2165,109 320.15 Wisconsin 1,9436,139 215.95 Wyoming 2,1185,812 174.41

SOURCE: U.S. Department of Education, National Center for Education Statistics, Digest of Education Statistics, 1994.(NCES 94-115). Table 166.


Table 2.—Total K–12 expenditures, by function: 1990–91

Percent of Percent ofExpenditures Total Current

Category (in thousands) Expenditures Expenditures Current Expenditures $201,549,624 88.04% Instruction 122,214,281 53.38 60.64% Student Services 70,419,509 30.76 34.94 Students 8,933,843 3.90 4.43 Instructional 8,467,453 3.70 4.20 General Administration 5,781,474 2.53 2.87 SchoolAdministration 11,680,254 5.10 5.80 Operation &Maintenance 21,323,871 9.31 10.58 Student Transportation 8,666,697 3.79 4.30 Other 5,565,916 2.43 2.76 Food Services 8,276,621 3.62 Enterprise Operations 639,213 0.28 Other Current Expenditures 3,298,439 1.44 Capital Outlay 19,770,913 8.64 Interest On School

District Debt 4,314,321 1.88 Total 228,933,297

SOURCE: U.S. Department of Education, National Center for Education Statistics, Digest of Education Statistics, 1993.(NCES 93-292). Table 159. Digest of Education Statistics, 1994. (NCES 94-115). Table 160.

expenditures per pupil, and in expendituresper pupil for direct instruction. This patternhas been confirmed by other researchers,most notably Bruce Cooper (1993, 1994).As discussed later in this paper, this lack ofvariation in the pattern of resource alloca-

tion may be part of the reason links betweenspending and student outcomes have beenhard to find, and may well offer possibilitiesfor making the allocation of resources moreproductive in the future.

Table 3.—Variation in per-pupil expenditures and the proportion of resources devoted toinstruction

Actual Cost Adjusted Per Pupil Per Pupil Per Pupil Per Pupil InstructionExpenditures Expenditures Expenditures Expenditures as a Percent for Current for for Current for of Total

Statistic Operations Instruction Operations Instruction ExpendituresMean $3,659 $2,137 $3,698 $2,194 $59.16Std. Dev. 1,912 961 1,759 825 6.28Coefficient ofVariation (%) 52.4 45.0 47.6 38.1 10.6

NOTE: These data are from the 1986–87 Census.SOURCE: Picus, L.O. 1993. “The Allocation and Use of Educational Resources: District Level Evidence from the

Schools and Staffing Survey.” Los Angeles, CA: The Finance Center of CPRE, Working Paper Number 34.


Trends in Student Outcomes

Despite this substantial increase inspending for education, many observers havepointed out that student outcomes have notimproved substantially, if at all. The resultsof standardized tests, international compari-sons of student performance, and othermeasures of the outcomes of schooling likeattendance rates and college enrollment havenot shown the same kind of increase as hasreal per pupil spending. However, manyargue that another important outcome ofschooling is success in the labor market, andsome have been able to link higher costresource allocation patterns to improvedcareer earnings. This section begins with adiscussion of student performance as typi-cally measured in education circles, focusingmost heavily on the results of standardizedtests. It concludes with a brief discussion oflabor market outcomes and their relation toschool spending.

Measures of Schooling. Despite thesubstantial increases in per pupil spendingobserved over the last 30 to 35 years, therehas not been a commensurate improvementin student performance as measured byscores on standardized tests. One of themost commonly used measures of studentperformance is the average ScholasticAssessment Test (SAT) score. Figure 1compares the average SAT score for highschool seniors with the change in real perpupil spending between 1968 and 1993.This familiar figure is often cited in discus-sions of why more money will not lead togreater student outcomes. Equally wellknown are the arguments that discredit thisanalysis. They include the increase in thenumber of students taking the SAT, thehigher numbers of minority students and thehigher number of students for whom Englishis not their first language who are taking theexam. As a result, some argue that the trendin SAT scores is not only expected, but thatit represents an improvement over the past.

Bracey (1994), for example, argues thatthe standards for the SAT were set in 1941based on the performance of less than11,000 students, 98 percent of whom werewhite, 60 percent of whom were male, andmost of whom lived in the Northeast.Moreover, a very high proportion attendedprivate high schools and intended to enrollin private colleges and universities. Braceygoes on to suggest that the bell curveimposed on this group led to only 6.68percent of them scoring above 650 on themathematics section of the SAT. Today,prior to the re-centering of the SAT scores,some 11 percent of the more than onemillion students taking the test score above650 on the mathematics section of the test,despite the fact that 30 percent are minority,52 percent female, and over 30 percentcome from families with incomes less than$30,000 per year. However, only 3.3percent score that high on the verbal portionof the test (Digest of Education Statistics,1994).

What is clear is that the SAT results donot provide a definitive answer to thequestion of whether or not the additionalmoney spent on education leads to improvedstudent outcomes. Moreover, because of thediversity in school districts, and the varia-tion in the percentage of students taking theSAT across states, it is not possible tocorrelate average state or district SAT scoreswith per pupil expenditures.

Another way to measure trends instudent performance is to review the resultsof the National Assessment of EducationalProgress (NAEP) over time. Mullis et al.(1994) show the following trends in perfor-mance for students age 9, 13, and 17 inscience, mathematics, reading, and writing:

• Science. Performance in sciencedeclined for all three age groups in the1970s, but improved during the 1980s. By1992, science performance at age 9 washigher than it was in 1969–70, while it was

The results ofstandardized

tests,international

comparisons of

studentperformaance,

and other

measures of theoutcomes of

schooling...have

not shown thesame kind of

increase as has

real per pupilspending.


lower for 17-year-olds and the same for 13-year-olds. Thus, only 9-year-olds per-formed significantly better in science in1992 than did 9-year-olds in 1969–70.

• Mathematics. Proficiency inmathematics improved between 1973 and1992 for those aged 9 and 13, while for 17-year-olds, proficiency declined in the 1970sand improved during the 1980s, returningto approximately the same level as observedin 1973.

• Reading. For 9-year-olds, readingperformance improved in the 1970s, de-clined in the 1980s, and by 1992, was atabout the same level as the first assessmentin 1971. Although there was little change inreading performance over time for 13-year-olds, over the 21 year period, averageperformance improved. For 17-year-olds,there were significant gains from 1971 to1984, although reading performance has

been flat since then.

• Writing. Assessed by grade levelrather than age, there has been little changein writing performance between 1984 and1992 for 4th and 11th graders. Fourthgraders showed a decline in 1990, and animprovement in 1992. On the other hand,8th graders showed a drop in performancebetween 1984 and 1990, with a significantimprovement by 1992. The size of the 8thgrade gain is so large that many are ques-tioning the results, and NAEP officials aretaking a wait-and-see approach and waitingfor additional assessments to be completedbefore jumping to any conclusions.

However one looks at these data, studentperformance in all four subject areas did notimprove at the 22 percent rate of increase inspending during the 1970s, nor at the 48percent increase of the 1980s.

1968

1970

1971

1972

1973

1974

1975

1976

1977

1978

1979

1980

1981

1982

1983

1984

1985

1986

1987

1988

1989

1990

1991

1992

1993

0

860

880

900

920

940

960

$0

$1,000

$2,000

$3,000

$4,000

$5,000

$6,000

Year

SAT

Sco

res

Per

Pupi

l Exp

endi

ture

s(1

993–

94 d

olla

rs)

Figure 1.—Trends in SAT scores and per pupil expenditures: 1968–1993

SAT scoreExp. Per Pupil

SOURCE: U. S. Department of Education, National Center for Education Statistics, Digest of Education Statistics, 1994. (NCES 94-115).

<>


To deal with[the] question,[Does money

matter ineducation?], it is

important to

consider whetheror not systematic

links between

educationalresources and

student outcomes

can be found.

There have also been a number ofinternational comparisons of the perfor-mance of U.S. students with students inother countries. Stevens and Stigler (1992)compared the performance of students inChicago and Minneapolis with similar agedstudents in Japan, Taiwan, and China. Theyshow how poorly our students do comparedto students in similar-sized cities in thesecountries. A number of other studies haveshown children in the United States perform-ing at lower levels than similar aged childrenin other countries.

These findings are somewhat controver-sial and highlight many of the difficultiesinvolved in comparing student achievementacross national borders. However, Bracey(1995) suggests that closer analysis ofinternational test results shows that studentsin the United States do not do as poorly aswe have been led to believe. He particularlypoints out that in the rankings of the SecondInternational Assessment of EducationalProgress, American 9-year-olds were rankedthird in the world in science. Although the14-year-olds were ranked 13th out of 15nations, Bracey goes on to point out that inboth cases the American students’ scoreswere very close to the average of the entiresample, and that the outcomes by countrywere very closely bunched.

Despite the continued debate over thecomparability of test scores over time in theUnited States, and across nations, it is clearthat measures of student performance havenot increased at the same rate as spending.In and of itself, this does not indicate thatmoney does not matter. To deal with thatquestion, it is important to consider whetheror not systematic links between educationalresources and student outcomes can befound. Before looking more closely atanalyses of this complex question, it ishelpful to consider the impact of educationalspending on labor market outcomes.

• Labor Market Outcomes. It canbe argued that an important outcome ofschooling is the ability of graduates to findand keep good, high paying jobs. Whilemaking the link between educationalresources and employment (as measured bylifetime earnings or a similar measure) isdifficult, on the basis of national data, Cardand Kruger (1990) found that men whowere educated in states with relatively smallclasses in the public schools and relativelyhigh teacher salaries tended to have higherearnings than did men educated in stateswith relatively larger classes and relativelylower paid teachers. Murnane (1991)suggests that Card and Kruger’s findingsmight lead to the conclusion that smallclasses and high teacher salaries (both ofwhich would lead to higher per pupilexpenditures), may have a greater effect onfuture earnings than on standardized tests.Murnane points out that this research canalso be challenged, and that there may beother factors leading to the correlationbetween small classes and highly paidteachers, and expresses concern that manystudies do not consider how expenditurelevels impact behavior of teachers andstudents. While more study will be requiredto answer this question definitively, it isclear that the long term impact of educa-tional spending decisions as measured bylabor market outcomes cannot be ignoredany more than can the short term resultsprovided by standardized tests.

The question that remains is whether ornot a systematic link between resources andstudent outcomes, no matter how defined,can be found. The next section of this papersummarizes the research that has been donein this area.

Production Function Analyses:What Have We Learned SoFar?


The analysis presented above leavessome doubt as to whether or not moneyreally matters in improving education. Thisis, and continues to be, a matter of consider-able debate today. Most of the research onthis issue focuses on production functionanalyses that attempt to relate educationalinputs to schooling outcomes. This sectionof the paper provides a brief description ofproduction functions, and discusses theiruse in educational research. This is fol-lowed by an analysis of the current debateover what existing research tells us aboutthe effect of resources on educationalattainment.

Production Functions: Their Use inEducation

A production function is a model thatidentifies the possible outcomes that can beachieved with a given combination ofinputs. With knowledge of the quantities ofinputs available, it is possible to calculatethe maximum output that can be achieved.What is important to this process is how theinputs are translated into those outcomes,and finding the most efficient way of doingso. The difficulty with identifying produc-tion functions in education results from thecomplexity of the schooling process and thenumber of inputs that can impact theoutcome. In addition, it is often difficult toreach agreement as to the desired outcomesof the educational system. Moreover, manyof the factors that appear to have an impacton the educational production process maywell be outside of the control of educators.

Production functions are estimatedthrough statistical or econometric tech-niques that rely on regression methods tomeasure the relationship between a mix ofinputs and some identified output. Amongthe most common outcomes used in studiesof educational production functions are theresults of standardized tests, graduationrates, dropout rates, and as discussed above,

labor market outcomes. The inputs mostoften considered include per pupil expendi-tures, pupil/teacher ratios, teacher education,experience and salary, school facilities, andadministrative inputs. Unfortunately, theresults of studies that have attempted tomeasure the effects of these inputs oftenconflict or show inconclusive results.Others, beginning with the well knownColeman Report (Coleman et al. 1966), haveshown that factors such as students’ socio-economic status may be more important indetermining how well they do in school thanare many, if not all of the inputs listed above.

Does Money Matter?: The CurrentDebate

While interest in the question of whetheror not money matters has always been high,the publication of an article by Hedges,Laine, and Greenwald (1994) in the April1994 Education Researcher has sparked arenewed debate over this issue. Prior to thepublication of their article, the most oftencited research in this field was the work ofEric Hanushek (1981, 1986, and 1989). Inan analysis of data from 38 different articlesand books containing 187 different regres-sion equations, Hanushek focused on theeffect of seven inputs to schooling. For eachinput, Hanushek analyzed the regressioncoefficients to determine if they werepositive or negative, and whether the effectmeasured was statistically significant. Heincluded a fifth category for those coeffi-cients that were not statistically significant,and for which the sign on the coefficientcould not be determined. His findings foreach of the seven inputs are summarized.

• Teacher/pupil ratio. Hanushekfound a total of 152 studies that consideredthe teacher/pupil ratio. While it is generallyaccepted that smaller classes lead to higherstudent achievement, of the 27 studies thathad statistically significant findings, only 14found that reducing the number of pupils perteacher was positively correlated to student

...factors such asstudents’ socio-

economic statusmay be moreimportant in

determining howwell they do inschool than...

many...inputs...


outcomes, while 13 found the oppositeeffect. Moreover, of the 125 that did nothave statistically significant results, 34found a positive effect, 46 a negative effect,and the sign or direction of the effect couldnot be determined for the remaining 45studies.

Interestingly, despite the lack of statisti-cally significant findings that lower pupil/teacher ratios lead to improved schoolperformance, there has been a dramaticeffort to reduce class size across the nationin the last 20 years. A cursory review of themost recent edition of the Digest of Educa-tion Statistics (NCES 1994) shows that theaverage pupil/teacher ratio for K–12 publicschools in the United States was 17.6:1 in1994. Moreover, the data provided in theDigest suggest that this ratio has declinedconsistently since 1955 when it stood at 26.9pupils per teacher (NCES, 1994, p. 74). Infact, except for an increase of 0.1 pupils perteacher between 1961 and 1962 and againbetween 1980 and 1981, and some minorincreases in the 1990s from the low of 17.2:1in 1989 and 1990, the average pupil/teacherratio across the United States has declined inevery year since 1955.

Picus (1993a) found that district levelpupil/teacher ratios declined as expendituresper pupil and expenditures per pupil forinstruction increased. However, as thepercent of expenditures devoted to instruc-tion increased, a similar pattern did notemerge. Since expenditure data were notavailable at the school level in the nationaldata bases, Picus (1993b) compared schoollevel pupil/teacher ratios with district perpupil expenditures. He found that at theelementary, intermediate, and secondaryschool levels, there is a trend toward lowerpupil/teacher ratios as expenditures increase.

• Teacher education. In looking atteacher educational attainment, Hanushekfound results similar to his findings for thepupil/teacher ratio. A total of 113 studies

considered teacher education. Hanushekfound that 8 of the 13 studies with statisti-cally significant coefficients showed apositive effect of teacher education onstudent performance. The statisticallyinsignificant studies were about evenlydivided among positive, negative, andindeterminate findings.

• Teacher experience. Hanushekfound more positive correlation betweenteacher experience and student performance.However, he indicated that these resultsonly appeared strong in relation to the otherfindings, and that they could be the result ofmore experienced teachers being able toselect teaching assignments with “goodstudents.” (Hanushek 1993).

• Teacher salary. Eleven of the 15studies Hanushek identified with statisti-cally significant results identified a positiverelationship between teacher salaries andstudent performance. However, he arguedthat even this was not particularly strongevidence of a relationship given that teachersalaries are largely determined by teachereducation and experience, and the underly-ing components of teacher salary wereunrelated to student performance.

• Per pupil expenditure. It is alsonot surprising that Hanushek found perpupil expenditures did not play an importantrole in determining student performance.Specifically, he found that 13 of 16 studieswith statistically significant results show apositive relationship. However, he dis-counted the importance of this arguing that8 of the 13 came from one study which hefelt did not measure family inputs precisely.As a result, Hanushek (1989) contends thatschool expenditures may have served as aproxy for family background in thosestudies.

• Administrative inputs. Monk(1989) points out that if there were not a

...despite the lackof statistically

significantfindings thatlower pupil/

teacher ratioslead to improved

school

performance,there has been adramatic effort

to reduce classsize...


Monk (1989)points out that ifthere were not a

productionfunction ineducation, then

the impact ofany input wouldnot matter to the

outcome.

production function in education, then theimpact of any input would not matter to theoutcome. Therefore, some amount ofadministration is essential to the operationof the educational system. Hanushek (1989)concluded that administrative inputs did nothave a systematic relationship to studentperformance, even though seven of the eightstudies with statistically significant findingsshowed a positive effect. Hanushek arguedthat variations in how administrative inputsare measured undermined the findings.

• Facilities. Adams (1994), Firestoneet al. (1994), and Picus (1994b) show thatwhen poor school districts in Kentucky,New Jersey, and Texas received substantialincreases in funding as a result of schoolfinance reforms in those states, a frequentresponse was to devote a substantial portionof those funds to facility improvements.This would imply that many educatorsbelieve the quality of school facilities areimportant to student learning. YetHanushek’s (1989) analysis of facilitiesshowed little relationship between thequality of school facilities and the perfor-mance of students.

After reviewing all 187 studies,Hanushek concluded that “There is nostrong or systematic relationship betweenschool expenditures and student perfor-mance.” (Hanushek 1989, 47). These wordshave been cited often by those opposed toproviding additional funds to the publicschools. Opponents of increased fundingargue that until educators can show moremoney will make a difference, additionalfunds should not be provided.

A recent review of these studies byHedges, Laine, and Greenwald (1994)questions Hanushek’s findings and suggeststhat money may in fact be more important indetermining how well students are likely todo. Hedges, Laine, and Greenwald (1994)reviewed the same studies Hanushekconsidered in his analysis. They eliminated

those studies that had insignificant resultsand the sign on the coefficient which couldnot be determined. They then analyzed theremaining studies which relied on statisticaltechniques other than the vote countingprocedure used by Hanushek. They con-cluded that

These analyses are persuasive inshowing that, with the possibleexception of facilities, there isevidence of statistically reliablerelations between educationalresource inputs and schooloutcomes, and that there is muchmore evidence of positiverelations than of negativerelations between resourceinputs and outcomes. (Hedges,Laine, and Greeenwald 1994,11)

While Hedges, Laine, and Greenwald(1994) found that expenditures do matter,they found less evidence of a relationshipbetween the other factors identified aboveand student performance. They suggestspecific allocation of those resources maynot be important in improving studentperformance in all situations. Further, theyargue that local authorities should be giventhe discretion to spend funds as they thinkwill best help the students for whom they areresponsible.

Hedges, Laine, and Greenwald (1994)point out that if, for example, per pupilexpenditures and student achievement wereunrelated, half the studies would havepositive coefficients and half negativecoefficients. Moreover, they argue if therewere no systematic relationship, only 5percent of the studies would have statisti-cally significant results. They then argue forthe studies, where the direction of thecoefficient could be determined, that ahigher percentage of the coefficients showeda positive sign. In fact, the three authorsargue this happened more often than would


be expected by chance alone.

Hedges, Laine, and Greenwald alsocriticize the vote counting method used byHanushek. They argue that as a procedure ithas limited power in finding significanteffects, and argue earlier work by Hedgesand Olkin (1980) shows that as the numberof studies reviewed increases, the probabilitythat a vote count will correctly detect aneffect decreases. Relying on a variety ofanalysis techniques, Hedges, Laine, andGreenwald (1994) re-analyzed Hanushek’sstudy sample, and concluded that expendi-tures do have an impact on student achieve-ment.

In a rejoinder, Hanushek (1994a) arguesthat the evidence still strongly indicates thatthere is no systematic relationship betweenhow much money is spent and how wellstudents perform. He asks, if money mat-ters, but class size and teacher characteristicsare not as important, what factors do in factmatter? Since teachers salaries and classsize are the two largest determinants ofspending, he suggests that if they are unim-portant, other components of spending mustbe more effective in improving studentperformance. He suggests few would agreewith the proposition that administration,another major component of school expendi-tures, is responsible for improving studentperformance.

In another study, Hanushek (1993)looked at the impact of spending on studentperformance in Alabama as measured by thepercent of students passing standardizedreading, mathematics, and language tests inthe third, sixth, and ninth grades. Despitethat fact that his analysis did not yieldstatistically significant results, he concludedthat if his estimates were used and spendingin each school district were increased to thelevel of the highest spending district in thestate (some $5,113 per pupil, at a cost of$1.05 billion above the current spending of$2.4 billion), student performance would

only be expected to improve by approxi-mately 4 percent, at most. In one instance,grade 6 language performance, Hanushekactually predicted that the increased spend-ing would reduce student performance by0.2 percent.

These are not the only studies that haveconsidered this question. A study byFerguson (1991) looked at spending and theuse of educational resources in Texas. Heconcluded that “hiring teachers with stron-ger literacy skills, hiring more teachers(when students-per-teacher exceed 18),retaining experienced teachers, and attract-ing more teachers with advanced trainingare all measures that produce higher testscores in exchange for more money.” (485)His findings also suggest that teachers’selection of districts in which they want toteach is affected by the education level ofthe adults in the community, the racialcomposition of that community, and thesalaries in other districts and alternativeoccupations. This implies, according toFerguson, that better teachers will tend tomove to districts with higher socioeconomiccharacteristics if salaries are equal. Ifteacher skills and knowledge have an impacton student achievement (and Ferguson, aswell as others suggest that it does) then lowsocio-economic areas may have to offersubstantially higher salaries to teachers toattract and retain high quality instructors.This would help confirm a link betweenexpenditures and student achievement.

One of the problems with all of thesestudies is they don’t take into considerationthe tremendous similarity with which schooldistricts spend the resources available tothem. As described above, research byPicus (1993a, 1993b, and 1994a) andCooper (1993, 1994) has shown resourceallocation patterns across school districts tobe remarkably similar, despite differences intotal per pupil spending, student characteris-tics, and district attributes. This does notmean that all children receive the same level

[Ferguson’s]findings...suggest

that teachers’selection of

districts in which

they want toteach is affected

by the education

level of theadults in the

community, the

racialcomposition of

that community,

and the salariesin other districts

and alternative

occupations.


If teacher skillsand knowledgehave an impact

on studentachievement...then low socio-

economic areasmay have to offersubstantially

higher salaries toteachers toattract and retain

high qualityinstructors.

of educational services. As Picus (1994a)points out, a district spending $10,000 perpupil and $6,000 per pupil for direct instruc-tion is able to offer smaller classes, betterpaid, and presumably higher quality, teach-ers, and higher quality instructional materi-als than is a district spending $5,000 perpupil and only $3,000 per pupil for directinstruction.

What we don’t know is what the impacton student performance would be if schoolsor school districts were to dramaticallychange the way they spend the resourcesavailable to them. In 1992, Odden andPicus suggested that the important messagefrom the research summarized above wasthat, “if additional education revenues arespent in the same way as current educationrevenues, student performance increases areunlikely to emerge.” (Odden and Picus1992, 281). Therefore, knowing whether ornot high performing schools utilize re-sources differently than other schools wouldbe very helpful in resolving the debate overwhether or not money matters.

Picus and Nakib (forthcoming) lookedat the allocation of educational resources byhigh performing high schools in Florida andcompared those allocation patterns with theway resources were used in the remaininghigh schools in that state. A total of sevendifferent measures were used to comparestudent performance. In preliminaryfindings, Picus and Nakib show per pupilspending and per pupil spending for instruc-tion was not statistically significantly higherin high performing high schools, largelybecause of the highly equalized schoolfunding formula used in Florida. On theother hand, they found the percent ofexpenditures devoted to instruction waslower in the high performing high schools,implying high performing high schools mayactually spend more money on resources notdirectly linked to instruction than do otherhigh schools.

Unfortunately, the results of this Floridaanalysis do little to clarify the debate onwhether or not money matters. Comparisonsof high performing high schools with allother high schools in Florida did not show aclear distinction in either the amount ofmoney available or in the way resources areused. As with many other studies, it wasstudent demographic characteristics that hadthe greatest impact on student performance.

Conclusion

This paper has shown that despiteconsiderable research on the matter, there isstill a great deal of debate as to whether ornot money makes a difference in education.Even though everyone agrees that highspending provides better opportunities forlearning, and seemingly higher studentachievement, statistical confirmation of thatbelief has been hard to develop. It is clearthat over the past 30 to 35 years, there havebeen dramatic increases in real per pupilrevenues for K–12 public education. De-spite the substantial cumulative increase, theannual average increase has averaged only2–3 percent. Consequently, educators andcommunity school boards have had littleopportunity to consider how they would uselarge increases in funding. As a result,educational resource allocation patterns areremarkably similar regardless of spendinglevel.

A careful look at the research on theimpact of money on student achievementshows that we may be asking the wrongquestion. Rather than consider whether ornot additional resources will improveeducational spending, it seems more impor-tant to ask how additional resources could bedirected to improve student learning, or inHanushek’s (1994b) view, spend thoseresources more efficiently.

Although the growth in educationalrevenues was flat in the early part of the


...it seems that ifour schools are

to succeed in the

future, it isimportant toprovide local

educators withthe resources and

tools they need

to meet thespecific needs ofthe children they

serve...

1990s, as the nation’s economy recoversfrom the recession, we are beginning to seeincreases in educational spending again.However, the mood of the public, while stillgenerous toward education, has changed.Now they seem to be insisting that schoolsshow dramatic improvements in exchangefor continued support. To meet this respon-sibility to both the taxpayers and schoolchildren, more research is needed to see ifthere are substantial differences in the wayhigh performing schools utilize their re-sources compared to other schools.

What seems evident today is that al-though aggregate analyses of school re-source allocation patterns leads to theconclusion that schools look remarkablyalike, the needs of the students within thoseschools are vastly different. The needs ofpoor, and often limited English speakingstudents in our inner cities, are considerablydifferent from those of middle and upperclass children in well-to-do suburbs acrossthe nation. Therefore, it seems that if ourschools are to succeed in the future, it isimportant to provide local educators with theresources and tools they need to meet thespecific needs of the children they serve, andat the same time allow them to designprograms that are specifically targeted to

those children.

If we can move away from measuringschool accountability through the way fundsare used, and instead measure accountabilityin terms of student outcomes, the answer tothe question posed in this paper will becomeunimportant. It won’t be whether or notmoney matters, but how that money is usedthat matters.


REFERENCES

Adams, J.E. 1994. “Spending School Reform Dollars in Kentucky: Familiar Patterns and New Programs, But IsThis Reform?” Educational Evaluation and Policy Analysis 16(4): 375-390.

Barro, S.M. 1992. “What Does the Education Dollar Buy?: Relationships of Staffing, Staff Characteristics, andStaff Salaries to State Per-Pupil Spending.” Los Angeles, CA: The Finance Center of CPRE, Working Paper.

Bracey, G.W. 1994. “The Fourth Bracey Report on the Condition of Public Education.” Phi Delta Kappan(October): 76(2): 114-127.

Bracey, G.W. 1995. “The Right’s Data-Proof Ideologues.” Education Week (January 25): XIV(18): 48, 37.

Card, D., and A. Kruger. 1990. “Does School Quality Matter?: Returns to Education and the Characteristics ofPublic Schools in the United States.” National Bureau of Economic Research, Working Paper No. 3358.

Clune, W. H. 1994. “The Shift from Equity to Adequacy in School Finance.” Educational Policy 8(4): 376-395.

Coleman, J.S., E.Q. Campbell, C.J. Hobson, J. McPartland, A.M. Mood, F.D. Weinfield, and R.L. York. 1966.Equality of Educational Opportunity. Washington, DC: U.S. Department of Health, Education and Welfare.

Cooper, B. 1993. “School-Site Cost Allocations: Testing a Micro-Financial Model in 23 Districts in Ten States.”Paper prepared for the Annual Meeting of the American Education Finance Association in March, 1993.

Cooper, B.S. 1994. “Making Money Matter in Education: A Micro-Financial Model for Determining School-Level Allocations, Efficiency, and Productivity.” Journal of Education Finance 20(1): 66-87.

Ferguson, R.F. 1991. “Paying for Public Education: New Evidence on How and Why Money Matters.” HarvardJournal on Legislation 28: 458-498.

Firestone, W.A., M.E. Goertz, B. Nagle, and M.F. Smelkinson. 1994. “Where Did the $800 Million Go? TheFirst Years of New Jersey’s Quality Education Act.” Educational Evaluation and Policy Analysis 16(4): 359-374.

Glass, G.V., and M.L. Smith. 1979. “Meta-Analysis of Research on Class Size and Achievement.” EducationalEvaluation and Policy Analysis January-February: 1(1): 2-16.

Greenwald, R., L.V. Hedges, and R.D. Laine. 1994. “When Reinventing the Wheel is Not Necessary: A CaseStudy in the Use of Meta-Analysis in Education Finance.” Journal of Education Finance 20(1): 1-20.

Hanushek, E.A. 1981. “Throwing Money at Schools.” Journal of Policy Analysis and Management 1: 19-41.

Hanushek, E.A. 1986. “The Economics of Schooling: Production and Efficiency in Public Schools.” Journal ofEconomic Literature 24: 1141-1177.

Hanushek, E.A. 1989. “The Impact of Differential Expenditures on School Performance.” Educational Re-searcher 18(4): 45-65.


Hanushek, E.A. 1991. “When School Finance ‘Reform’ May Not be a Good Policy.” Harvard Journal onLegislation 28: 423-456.

Hanushek, E.A. 1993. “Can Equity be Separated from Efficiency in School Finance Debates?” Essays on theEconomics of Education W.E. Upjohn Institute for Employment Research. Ed. E.P. Hoffman. Kalamazoo, MI.

Hanushek, E.A. 1994b. Making Schools Work: Improving Performance and Controlling Costs. Washington,DC: The Brookings Institution.

Hanushek, E.A., and S.G. Rivkin. 1994. “Understanding the 20th Century Explosion in U.S. School Costs.”Rochester, NY: Rochester Center for Economic Research, Working Paper No. 388.

Hanushek, E.A. 1994a. “Money Might Matter Somewhere: A Response to Hedges, Laine, and Greenwald.”Educational Researcher 23(4): 5-8.

Hedges, L.V., R.D. Laine, and R. Greenwald. 1994. “Does Money Matter? A Meta-Analysis of Studies of theEffects of Differential School Inputs on Student Outcomes.” Educational Researcher 23(3): 5-14.

Hedges, L.V., and I. Olkin. 1980. “Vote Counting Methods in Research Synthesis.” Psychological Bulletin88: 359-369.

Laine, R.D., R. Greenwald, and L.V. Hedges. 1994. “The Use of Global Education Indicators: Lessons LearnedFrom Education Production Functions, Lesson #1: Money Matters.” Paper presented at the annual meeting ofthe American Education Research Association on April 4, 1994 in New Orleans, LA.

Marsh, D.D., and J.M. Sevilla. 1992. “Financing Middle School Reform: Linking costs and Education Goals.”in Rethinking School Finance: An Agenda for the 1990s. Ed. A.R. Odden. San Francisco, CA: Jossey-Bass.pp. 97-127.

Monk, D.H. 1989. “The Education Production Functions: Its Evolving Role in Policy Analysis.” EducationalEvaluation and Policy Analysis 11(1): 31-45.

Monk, D. 1992. “Educational Productivity Research: An Update and Assessment of Its Role in EducationFinance Reform.” Educational Evaluation and Policy Analysis 14(4): 307-332.

Mullis, I.V.S., J.A. Dossey, J.R. Campbell, C.A. Gentile, C. O’Sullivan, and A.S. Latham. 1994. Report in Brief:NAEP 1992 Trends in Academic Progress Achievement of U.S. Students in Science, 1969 to 1992, Mathematics,1973 to 1992, Reading, 1971 to 1992, Writing, 1984 to 1992. Washington, DC: U.S. Government PrintingOffice. Report Number 23-TR01.

Murnane, R.J. 1991. “Interpreting the Evidence on “Does Money Matter?” Harvard Journal on Legislation28: 457-464.

Nakib, Yasser. 1995. Resource Allocation and Student Achievement at the School Level: What Seems to Matter?Los Angeles, CA: Center for Research in Education Finance.

U.S. Department of Education, National Center for Education Statistics, Digest of Education Statistics, 1994.NCES 94-115. Washington, DC.


Odden, A. 1990. “Class Size and Student Achievement: Research Based Policy Alternatives.” EducationalEvaluation and Policy Analysis (Summer 1990): 12(2): 213-227.

Odden, A. 1994. “The Local Impact of School Finance Reform in Kentucky, New Jersey and Texas.” Educa-tional Evaluation and Policy Analysis 16(4).

Odden, A.R., and L.O. Picus. 1992. School Finance: A Policy Perspective. McGraw-Hill. New York, NY.

Picus, L.O. 1993a. “The Allocation and Use of Educational Resources: District Level Evidence from theSchools and Staffing Survey.” Los Angeles, CA: The Finance Center of CPRE, Working Paper Number 34.

Picus, L.O. 1993b. “The Allocation and Use of Educational Resources: School Level Evidence from theSchools and Staffing Survey.” Los Angeles, CA: The Finance Center of CPRE, Working Paper Number 37.

Picus, L.O. 1994a. “The $300 Billion Question: How Do Public Elementary and Secondary Schools SpendTheir Money?” Paper presented at the 1994 Annual Meeting of the American Education Research Association.New Orleans, LA.

Picus, L.O. 1994b. “The Local Impact of School Finance Reform in Four Texas School Districts.” EducationalEvaluation and Policy Analysis 16(4): 391-404.

Picus, L.O. and Y. Nakib. Forthcoming. “Resource Allocation Patterns in High Performing Florida Schools.”Los Angeles, CA: Center for Research in Education Finance.

Stevens, H.W., and J.W. Stigler. 1992. The Learning Gap. Simon and Schuster. New York, NY.

Word, E., J. Johnston, H.P. Bain, D. Fulton, J.B. Zaharies, M.N. Lintz, C.M. Achilles, J. Folger, and C. Breda.1990. Student/Teacher Achievement Ratio (STAR), Tennessee’s K-3 Class Size Study: Final Summary Report,1985-1990. Nashville, TN: Tennessee State Department of Education.

The Effect of ConstitutionalLitigation on Educational Finance:

A Further Analysis

Selected

Papers in

School

Finance


G. Alan Hickrod is DistinguishedProfessor Emeritus of Educational Adminis-tration and Foundations at Illinois StateUniversity (ISU). Dr. Hickrod founded theCenter for the Study of Education Financeat ISU which has explored educationalfinance topics for over two decades. Hewas also a founding member of the Ameri-can Education Finance Association, was itstenth President, and holds its distinguishedservice award.

Dr. Hickrod has directed approximately18 different educational research grants andauthored numerous publications for aca-demic review. His work includes over 75monographs and books and over 65 articlesand papers on school finance and education

research. Many of his publications haveregularly appeared in the Journal of Educa-tion Finance, Journal of EducationalAdministration, Educational AdministrationQuarterly, American Education ResearchJournal, Review of Educational Research,and many more. He is also a regular con-tributor to the “Constitutional Challenges”and “Selected Readings” sections for theIllinois State Board of Education publicationtitled, State, Local, and Federal Financingfor Illinois Public Schools.

Dr. Hickrod is a Harvard graduate with adoctorate in education. He is a leadingauthority on school finance, social andeconomic foundations, and statistics andresearch methods.

Ramesh Chaudhari is Coordinator ofTechnology Transfer for the College ofEducation at ISU and is also a ResearchAssociate at the Center for the Study ofEducation Finance.

The Effect of ConstitutionalLitigation on Educational

Finance: A Further Analysis

G. Alan HickrodRamesh Chaudhari, Gwen Pruyne, Jin Meng

Illinois State University

About the Authors

Gwen Pruyne is a Research Associate atthe Center.

Jin Meng served as Research Assistantat the Center for this project.

The Effect of Constitutional Litigation 39

Background and Prior Studies

The late Charles Scott Benson, testify-ing before a U.S. Senate Committee not longago said, “You must be very careful whenyou wish for things because you may justget what you wish for. We worked hard forequity in California. We got it. Now wedon’t like it.” Before Professor Bensonmade his sage observation, James GordonWard, in a seminal conceptual piece hadnoted that the goals of “equity” and “ad-equacy” might well be in conflict (1990).

Despite the timely warnings by Bensonand Ward, however, 44 states have experi-enced litigation challenging the constitution-ality of their K–12 finance systems. Appen-dix A is a current update of the status ofthose state constitutional challenges. Forthe last four years, the American EducationFinance Association has conducted a pre-conference workshop for plaintiffs’ lawyersin these school finance constitutionalchallenges. Every year, the filing of newcomplaints has been reported in that work-shop. Obviously, a considerable number ofpeople think that something can be gainedby bringing these cases to court.

The Effect of ConstitutionalLitigation on EducationalFinance: A Further Analysis

G. Alan HickrodRamesh Chaudhari, Gwen Pruyne, Jin MengIllinois State University

Six studies were used in the empiricalquantitative research reported here. Theseeither bore directly on the question of theeffect of these constitutional challenges onschool finance, or provided the necessarydata for an investigation of that topic. Thefirst study was by the senior author of thisreport and his colleagues at the Center forthe Study of Education Finance, IllinoisState University. Hickrod et al. (1992) useda twenty-year time span (1970–90). Thisresearch group found that winning the caseat the state supreme court level had only amodest effect on increases in expenditureper pupil. However, winning at the statesupreme court level brought about a mean-ingful shift in tax burden from local to statesources, while losing at the state supremecourt level brought about the oppositeeffect—a shift from state to local resources.Winning the case was thus seen more as ameans of property tax relief than as a primedeterminate of more adequate funding.Winning also appeared to have little effecton fiscal effort for education. Using re-worked data from a study by James Wyckoff(1992), Hickrod et al. also concluded there

Obviously, aconsiderable

number ofpeople think thatsomething can be

gained bybringing thesecases to court.


tions have stimulated growth in expenditureper pupil in those twelve states with suc-cessful litigation.

There is a major difference in theresearch designs of Heise and Peternickfrom that used by the Hickrod group at ISU.While Heise and Peternick sought todetermine the effect of a decision at onepoint in time, the work at ISU assumes thatthe effect of decisions, note the plural, iscumulative over time. The Hickrod designis based on a notion that the litigation bringspressure upon the state legislature. Howmuch pressure can be brought by thelitigation on the state legislature? There issome evidence that the longer the court caseis pursued, the more the pressure on thelegislature builds. Winning, or winningfollowed by compliance litigation, puts themost pressure on the state legislature.Losing puts the least amount of pressure onthe legislature, but losing and refiling at alater date on different legal grounds stillapplies some pressure on the state legisla-ture. The Heise/Peternick approach is asingle sharp action while the Hickrodapproach is one of long term “climate” in astate. Neither design is wrong; they are justdifferent.

In addition to these empirical studies,the work by Hertert (1994), Riddle (1994),and Moskowitz, provided essential calcula-tions for the investigation reported here.1

These calculations of the amount of dispar-ity in each state, often measured as thecoefficient of variation, have made a veryvaluable contribution to the study of educa-tional finance. So much of the schoolfinance data of the past has been based onindividual state case studies that it has beendifficult to mount generalizations which willhold over all 50 states. This is not to

was some evidence to support the hypothesisthat winning, or winning and then filingcompliance litigation, had the effect ofreducing disparities between school districts.A disadvantage of the Wyckoff data was thelimited time period, 1980 to l987.

To date, the most sophisticated investi-gation of the question of the effect ofjudicial intervention is by Michael Heise(Fall, 1995). Heise used data from l970 tol992 to explore the effect of state supremecourt decisions in two states, Wyoming andConnecticut. The study is multivariate indesign and uses a dummy variable to repre-sent the decision at one point in time. Theresearch design is a single interrupted timeseries. Heise finds that in neither Wyomingnor Connecticut was the state supreme courtdecision statistically significant in increasingthe expenditure per pupil of those states.Other factors in his multivariate equationplayed a greater role in increasing expendi-ture per pupil than did the court decision.

A study by Lauri Peternick (1995) usedthe longitudinal financial record of 12 stateswhich had found their school financesystems to be unconstitutional (Arkansas,California, Connecticut, Kansas, Kentucky,Montana, New Jersey, Texas, Washington,West Virginia, Wisconsin, Wyoming). Bycomparing these states in which the systemwas adjudged unconstitutional with states inwhich the system was sustained, she soughtevidence on the question of whether adecision was detrimental to increases inexpenditure per pupil. While the Peternickstudy is not multivariate in nature, it isperhaps closer to the Benson-Ward notionthat seeking the goal of equity might bedetrimental to adequacy. Peternick’s re-search did not support that notion. To thecontrary, she found that the judicial interven-

1 Through the good offices of Stephanie Stullich, data from an unpublished study by Jay Moskowitz (commissioned by theU.S. Department of Education) was made available for the purpose of this investigation. Specifically, the coefficients ofvariation on state and local revenues per pupil for each state, 1980 and 1992, were used to compute percentage change inthose coefficients. Since the most important findings in this report hinge upon the changes in those coefficients, theauthors are properly appreciative of the efforts of both Stullich and Moskowitz.

While Heise andPeternick soughtto determine the

effect of adecision at onepoint in time,

the work at ISUassumes that the

effect of

decisions...iscumulative over

time.


deprecate individual state case studies.Given the de-centralized organization ofschool finance in the United States, thoseindividual state case studies will continue toplay an important policy role in Americanschool finance (Hickrod et al.).

Empirical Section

As in the previously reported study, theinvestigators assumed that the six categoriesof states pictured in Figure 1 constituted thesix different “climates” in which schoolfinance fiscal policy changes have takenplace. The first eight states of Category Iare those in which plaintiffs have won aclear judicial victory. In these states, it wasassumed that the pressure on the legislaturewould be at the greatest level. The six statesof Category II also constitute those in whichjudicial activity would have put greatpressure on the legislatures. In these states,there has been a clear judicial victory for theplaintiffs, but the plaintiffs have beenunhappy with the actions of the state legisla-ture and have sought additional relief fromthe courts. The 11 states of Category IIIshould exert less pressure on the legislaturessince in those states the defendant (the state)won and no further complaint has been filed,or, if a further complaint has been filed, thatcomplaint also lost. Less pressure shouldalso be found in Category IV, although thesesix states constitute something of an enigma.They are states in which a previous effort atlitigation has lost and yet plaintiffs have hadthe perseverance and stamina to file againon a different legal theory and perhaps evenfile a second or third major challenge. Thus,Category IV might indicate more pressurethan Category III.

In the ten states in Category V, thereshould be less pressure on the state legisla-ture since no state supreme court ruling hasyet taken place. There have been middlelevel court decisions in some of these states,in addition to lower court decisions. Insome of these states most of the action has

been on a motion to dismiss by the state,with no trial on the merits having yet beenheld. It would seem that the pressure on thelegislature brought about by an actual trialon the merits could be an important part ofthe determinants of fiscal policy which thisstudy wanted to investigate. A trial ofseveral weeks, with its attendant presscoverage, might well generate actions in thelegislature that nothing else could bringabout. Finally, Category VI states surelyconstitute the least pressure on the legisla-ture. In fact, in six of these nine states therehas been no litigation at all and in the otherfour states there has been a history oflitigation but the case is presently dormant.

It will be obvious to those familiar withthis body of litigation that this “climate”system is far from perfect. Assigning a statelike Kansas, where there has been a gooddeal of litigation activity, to Category VI isespecially doubtful. Also assuming that theaction at the lower court level in Alabamawas equal to the actions in the lower andmiddle level courts of Illinois is unfounded.Neither is there a comparison between thesetwo states with a very unusual supreme courtaction in Missouri. It may also be difficultfor plaintiff lawyers to accept that thedecision in North Dakota was actually a loss.However, even though the categories demon-strate some heterogeneity within class, thecategories are sufficiently distinct to beuseful for analysis. Both the previouslyreported work and the work reported hereconfirm that belief.

The first evidence is in Table 1 whichtakes four fiscal variables at one point intime and categorizes them by the six state“climates.” This is probably the weakest ofthe evidence reported here since manyvariables other than court decisions couldand do affect the magnitude of these fourfiscal variables. Nevertheless, there aresome interesting possibilities. The highestexpenditure levels are found in Categories IIand IV. Repeated litigation is found in both

It would seemthat the pressureon the legislature

brought about byan actual trial onthe merits could

be an importantpart of thedeterminants of

fiscal policy...


WASHINGTONMONTANA

CALIFORNIA

OREGON

NEW MEXICO

UTAH

OKLAHOMA

IDAHO

NEVADA

COLORADO

WYOMING

NORTH DAKOTA

SOUTH DAKOTA

NEBRASKA

ARIZONA

MINNESOTA

ILLINOIS

ARKANSAS

IOWA

TEXAS

KANSAS

MISSOURI

WISCONSIN

MICHIGAN

OHIO

KENTUCKY

ALABAMA

TENNESSEE

GEORGIA

MAINE

NEW YORK

N. CAROLINA

S. CAROLINA

NH

VT

MA

CT

NJ

DEMD

RI

INDIANA

MISSISSIPPI

LOUISIANA

FLORIDA

II

IIVI

VI

I

V

V

III

III

I

VI V

IIIIII

VI

V

IV

III

IV

IV

III

V

V

I

VI

III

III

ALASKA

HAWAII

V

VI

IV

I

VI II

III

III

III

I

II

I

IV

VI V

VW. VA

IIVIRGINIA

PENNSYLVANIA

MARI CTNJDEMD

I V I II VI IV

-- ----

N

S

W E

DC

I. Plaintiffs won at the state supreme court level. (8 states)

II. Plaintiffs won at the state supreme court level, but further compliance litigation was filed. (6 states)

III. Plaintiffs lost at the state supreme court level and there has been no further complaint filed or lost. (11 states)

IV. Plaintiffs lost at the state supreme court level, but there have been further complaints filed. (6 states)

V. Litigation was present, but no supreme court decision has been rendered. (10 states)

VI. No litigation is present or case is dormant. (9 states)

Figure 1.—Status of School Finance Constitutional Litigation

SOURCE: Center for the Study of Educational Finance, Illinois State University. Normal: Illinois.


of these categories. In Category II there hasbeen a win and then further litigation. InCategory IV there has been a loss and thenfurther litigation. The lowest expenditurelevels are found in the states that haveexperienced no litigation or where thelitigation is dormant.

With regard to the crucial matter ofdisparity between districts (Table 1, coeffi-cient of variation), at first there does notappear to have been much effect caused bycourt activity since the disparity for thewinning states is actually greater than thedisparity in the losing states. However, thispicture greatly changes when the focusbecomes change through time which is thesubject of Table 2. In Table 1, fiscal effort,defined as the ratio of expenditure per pupilto income per capita as of l993, also does notseem to have been effected much by courtdecisions. Percentage of state aid is high forCategory II (winning plus additional compli-ance litigation), but it is also high forCategory VI which has little litigationactivity. Again, this is clarified by Table 2.

The evidence of Table 2 is more sub-stantial. Here are arrayed the same fourfiscal variables, but now the focus is onchange through time in these four variables.

States in which there have been winningsupreme court decisions have a greaterpercentage increase than any other legalcategory. This finding would support thePeternick results. However, since nocontrols are effected here for other variablesthat are changing through time, as in theHeise study, one cannot be sure the increasein funding observed is due entirely to thecourt decision(s). The change in disparitybetween school districts within a state ismuch more prominent than is the increase inexpenditure per pupil. Here Category I(winning) states have clear reductions indisparity between school districts and thereis also a reduction in disparity in Category IIstates (win but with further compliancelitigation). By contrast, Category III states(losing states) show no decease in disparitybetween school districts and Category IVstates (loss but with further litigation) showstriking increases in disparity among schooldistricts. This makes some practical sense.It is quite likely that the increasinglydisparate situations in Ohio, Pennsylvania,and New York have driven the equityproponents in those states back into addi-tional litigation, despite some early losses inthat arena.

Table 1.—Means of selected fiscal variables, by legal category

CategoryFiscal Variables I II III IV V VI

Expenditure Per Pupil (1993) $5,463 $5,620 $5,000 $5,759 $5,034 $4,853

Coefficient of Variation (1990) 19.32 18.98 18.76 19.6020.70 15.36Fiscal Effort (1993) 0.2572 0.2834 0.2529 0.2739 0.2584 0.2512

Percentage State Aid (1990) 44.84 54.33 39.69 47.03 50.10 54.57

NOTE: Category I: plaintiffs won. Category II: plaintiffs won, further litigation filed. Category III: plaintiffslost, no further litigation filed. Category IV: plaintiffs lost, further litigation filed. Category V: no court decision.Category VI: no litigation present.

SOURCE: Derived from Common Core of Data, Fiscal F-33 Collection, U.S. Department of Education,National Center for Education Statistics, Washington, DC.

States in whichthere have been

winning supremecourt decisionshave a greater

percentageincrease [infunding] than

any other legalcategory.


support education. If the state governmentshoulders this responsibility, fine. However,in many states the legislature has beenreluctant to put more funds into publiceducation, or cannot put more funds intopublic education due to the increasingcompetition for those funds from functionssuch as Medicaid and various welfareprograms. Should the federal governmentpush more of these medical and welfareprograms back to the state level, then statefunds would be in even shorter supply foreducational programs. It might be evenmore difficult for the state to assume theresponsibility once held by the local schooldistrict. It is also possible that the“balkanization” of the state legislaturescould lead to even more “gridlock” betweenwealthy and poor parts of the state with theresult that no legislation is able to be passedto increase funding for the schools.

This is a rather pessimistic hypothesis,and essentially assumes that one cannotmake gains on both “adequacy” and “equity”at the same time. To explore this hypothesispercentage change in expenditure per pupil

Differences in changes in fiscal effortare not remarkable except that it is interest-ing to note that increases in fiscal effort arelowest in Category VI, the category with theleast legal activity. With respect to changein percentage of state aid provided to localschool districts, it is apparent that very largeincreases in state aid occurred in Category IIstates (plaintiff won and then filed compli-ance litigation after the win). This suggestswinning the lawsuit may bring about taxrelief to property taxpayers, a point made inthe ISU Center’s original investigation.

From the data assembled here, it ispossible to erect a test of the Benson/Wardhypothesis, at least a partial test. Essen-tially, the warning sounded by both Bensonand Ward is that, if the state puts all of itsefforts into reducing disparity betweenschool districts, then the overall state/localfunding level of the state for public educa-tion could either fall over time or remainstatic over time. This could come about inseveral ways. Reducing disparity requires agreater and greater reliance on the state,rather than the local unit of government, to

Table 2.—Change in means of selected fiscal variables, by legal category

CategoryFiscal Variables I II III IV V VI

Percentage Change of Expenditure $279 $255 $233 $259 $232 $236Per Pupil (1978–1993)

Percentage Change of Coefficient -22.14 -7.54 2.44 22.51 6.15 -5.73of Variation (1980–1990)

Percentage Change of Fiscal 1.37 1.4 1.34 1.45 1.37 1.23Effort (1980–1990)

Change of Percentage State Aid 3.44 19.25 5.36 1.73 5.69 7.14(1970–1990)

NOTE: Category I: plaintiffs won. Category II: plaintiffs won, further litigation filed. Category III:plaintiffs lost, no further litigation filed. Category IV: plaintiffs lost, further litigation filed. Category V: nocourt decision. Category VI: no litigation present.

SOURCE: Derived from Common Core of Data, Fiscal F-33 Collection, U.S. Department of Education,National Center for Education Statistics, Washington, DC.

Reducingdisparity requires

a greater andgreater reliance

on the state,

rather than thelocal unit of

government, to

supporteducation.


(1978–1993) was regressed upon percentagechange in the coefficient of variation ofexpenditures per pupil within the severalstates. This last dimension is a little confus-ing since there are negative percentagechanges (reductions in disparity) at thelower end of the scale and positive percent-age changes (increases in disparity) at theupper end of the scale. For the Benson/Wardhypothesis to hold, one would need to find apositive relationship between the twovariables. That is, low increases in expendi-ture per pupil would need to be related toreductions in disparity and high increaseswould need to be related to increases indisparity. However, the data we haveindicates not a positive, but rather a negativerelationship. High increases in expenditureper pupil are related to reductions in dispar-ity and low increases in expenditure perpupil are related to increases in disparitybetween school districts. Granted therelationship is not strong, being merely aPearson product-moment correlation of - .l5.A more cautious finding might be that nosignificant relationship exists between the“adequacy” goal, (increasing expenditures)and the “equity” goal (decreasing disparity).Since they do not seem to be related, “ad-equacy” and “equity” may well have differ-ent determinants.

Figures 2 and 3 constitute a Chi Squaretest of the significance of the state supremecourt decisions. In Figure 2 winning andlosing decisions are arrayed against high andlow increases in percentage change inexpenditure per pupil. There appears to belittle effect of the decisions upon the per-centage change in expenditure per pupil.The contingency coefficient is only .12 andnot statistically significant. By contrast,when the winning and losing decisions arearrayed against decreases in disparity or, atbest, small increases in disparity and largeincreases in disparity, there is an easilyascertained relationship. Winning decisionsare associated with reduction of disparityand losing decisions are associated with

increasing disparity. The relationship asstated by the contingency coefficient is .38which is significant at the .02 level. Thus,the evidence in this study supports thenotion that the state supreme court decisionshelp to reduce expenditure disparity be-tween school districts within states, but arenot necessarily related to increases inaverage expenditures per pupil in thosestates.

Anecdotal Evidence

In addition to the statistical analysis,anecdotal evidence was gathered to shedmore light on the claim that litigation hasbeen effective in obtaining increasedfinancial support for a state’s schools. Tobe included in this part of the study, a stateneeded to meet two criteria. First, it neededto be one of the 14 states in Categories I andII on the status report (Appendix A). Sincethe data analysis included the 1990s, it wasdecided that, with the passage of time,intervening conditions might have adverselyaffected the earlier gains, so the surveyshould be directed toward the more recentcases, thereby eliminating California(1977), Connecticut (1977), Washington(1978), and Wyoming (1980). The secondcriterion was that the statistical analysisgave evidence of either an increase in perpupil spending, a reduction in disparitiesbetween school districts, or both conditions.Table 3 lists the rankings among the 11states in which anecdotal materials werecollected.

Since the data placed Connecticutamong the highest on both criteria, it wasincluded in the interviews. The influence ofthe landmark decisions in California arereflected in the later cases in the analysis, soa brief comment about Serrano is included.

CALIFORNIA: The impact of Serranov. Priest (1971) in the early 1970s has had aprofound effect upon school fundingsystems in many states besides California.

...state supremecourt decisions

help to reduceexpendituredisparity between

school districtswithin states, butare not

necessarilyrelated toincreases in

averageexpenditures perpupil in those

states.


Figure 2.—Win or loss at supreme court level and percentage change in expenditure per pupil (1978–1993)

Percentage Change inExpenditure Per Pupil

Count High Low RowExp Val Increase Increase Total

Won 6 8 14Legal 5.1 8.9 46.7%Categories

Lost 5 11 165.9 10.1 53.3%

Column 11 19 30Total 36.7% 63.3% 100.0%

ApproximateStatistic Value Significance

Contigency Coefficient 0.11931 0.51043 *

* Pearson chi-square probability

SOURCE: Derived from Common Core of Data, Fiscal F-33 Collection, U.S. Department of Education, National Center forEducation Statistics, Washington, DC.

Figure 3.—Win or loss at supreme court level and percentage change in coefficient of variation

Percentage Change inCoefficient of Variation

DecreaseCount or Low Large RowExp Val Increase Increase Total

Won 11 3 14Legal 7.9 6.1 46.7%Categories

Lost 6 10 16 9.1 6.9 53.3%

Column 17 13Total 56.7% 43.3% 100.0%

ApproximateStatistic Value Significance

Contigency Coefficient 0.38211 0.02353 *

* Pearson chi-square probability

SOURCE: Derived from Common Core of Data, Fiscal F-33 Collection, U.S. Department of Education, National Center forEducation Statistics, Washington, DC.


During the so-called “first wave” of litiga-tion, even states which did not have lawsuitspending were awakened to the fact that theirfinance systems might be in jeopardy andthere was an upsurge of reform. TheSerrano dicta (which essentially stated thateducational funding was a function of thewealth of the state, not the wealth of adistrict) became an important part of manyplaintiffs’ cases. In the earlier suits, how-ever, the judiciary played a less active rolein developing guidelines or deadlines whichwere incumbent upon the legislatures.

ARKANSAS: In the fall of 1994, thecourt declared the state’s school-fundingformula unconstitutional and the legislaturewas given two years to correct it. Based onpast experience, some feared that, withoutmore guidance from the court, inequitieswould not be remedied. After the financesystem was ruled unconstitutional in 1978,the legislature enacted revisions to thestate’s complex formula. Despite thesechanges, financial disparities actually rosebetween 1979 and 1993. In the more recentdecision, the court was more directive inexpediting its rulings.

ARIZONA: A headline in EducationWeek stated, “Judge orders lawmakers to fixArizona finance system.” The reliance onproperty tax to pay for education was foundto be unconstitutional because it producesdisparities, especially in the qualities ofschool facilities. Several remedies havebeen suggested, but the finance issue wasdelayed until the 1995 session of the legisla-ture.

CONNECTICUT: The threat of litiga-tion in other states in the early 1970s mayhave caused the legislature to “start the ballrolling” on school finance reform, but thesupreme court decision in Horton v. Meskill(1977) brought it to a climax. The flat grantsystem was dropped and an equalizationformula was initiated. The formula took intoaccount the personal incomes of the variouscommunities and targeted money towardthose where the need was greatest, therebydecreasing the disparity between districts.The strong economy of the state in the 1980sfostered increased funding. When theeconomy “took a nose dive” in the early1990s, the legislature was committed toincreased funding. A personal income taxwas enacted, the sales taxes were decreased,but the property taxes remained about thesame. The continued strong support, in spite

Table 3.—Rankings for 11 states in which anecdotal materials were collected

Reduction in Increased Expenditure States Disparity1 Rank Per Pupil2 Rank Arizona -23.81 7 189.21 10 Arkansas -5.88 8 219.78 8 Connecticut -36.84 3 300.32 6 Kentucky -44.44 1 303.46 5 Massachusetts -4.55 9 219.33 9 Montana -25.00 6 172.25 11 New Hampshire -26.09 5 409.60 1 New Jersey -23.81 7 348.15 3 Tennessee 11.11 10 253.66 7 Texas -39.13 2 326.86 4 West Virginia -30.77 4 384.25 2

NOTE: Arizona and New Jersey are both ranked 7. 1 A negative number indicates a decrease in disparity. The higher the number, the smaller the disparity. 2

The higher the number, the greater the increase in expenditure per pupil from 1970 to 1990.

The KentuckyEducation

Reform Act hasbeen called the“national model

of schoolreform.”


districts to spend between 80 percent and100 percent of that level. To meet thisrequirement, some districts are restrictedfrom spending more than the standard andsome districts will be forced to increaseproperty taxes.

NEW HAMPSHIRE: Historically, thisstate’s role in education has been limited.The court ruled that education is a statefunction. While there have been proposalsto increase the level of state support, thelegislature has been slow to respond in partbecause the state has no sales or incometaxes and the political climate is un-support-ive of enacting new taxes. One possibleoption is to collect and distribute someproperty taxes on a statewide basis. Thecourt remanded the tax-equity question tothe trial court.

NEW JERSEY: Repeatedly, the Su-preme Court has ordered the legislature toremedy the inequities among its schooldistricts. This failure to implement a fund-ing reform reflects the “complexity andpolitical contentiousness of school financein New Jersey, where the issue has longpreoccupied state lawmakers and judges.”(Education Week, April 6, 1994) In 1994,the finance system was ruled unconstitu-tional for the third time. This time, thecourt specified that the legislature mustmake substantial equivalencies in per pupilspending by the 1997–98 school year.

TENNESSEE: The court providedgeneral guidelines and left the details to thelegislature. The Basic Education Programand a half cent sales tax increase wereenacted in 1992. There has been a sched-uled increase of 15 to 30 percent in overallstate funding which will be phased in over aperiod of six years. The formula requireslittle additional effort from local govern-ment; however, many local governmentshave increased their support. The reduceddisparity between high and low spendingdistricts has been noticeable, but the major

of an economic slump, might be attributed tothe somewhat unique political climate in thisstate which has been described as its “con-sensual nature of politics.”

KENTUCKY: The Kentucky EducationReform Act has been called the “nationalmodel of school reform.” The decision inKentucky was the beginning of a trend inwhich courts began playing a more promi-nent role in school finance litigation andlegislative solutions. As a result of thecourt’s findings, a new formula was estab-lished. Each district was guaranteed aspecified minimum amount of money perpupil. State funding was provided formandated programs. Local districts mustcontribute a fair share by taxing at a speci-fied minimum rate and may raise additionalfunds, with matching state funds provided insome situations. While those elements ofthe reform which deal with other than fiscalmatters are still debated, there is no questionthat the funding increased. Of the statessurveyed, Kentucky ranks first on reductionof disparities and among the highest inincreases in per pupil expenditure.

MASSACHUSETTS: The SupremeCourt’s ruling emphasized that the state hasprimary responsibility for education. Theeducation reform plan sets a foundation levelwhich is to be reached by the year 2000.The state will play a greater role in makingup for differences between wealthy and poordistricts and the state has specified a mini-mum amount that local governments mustappropriate to participate in the state’seducation funding system.

MONTANA: In 1989, the SupremeCourt found, among other things, that thefunding system was forcing an over-relianceon property tax levies. The legislatureresponded by rewriting the funding system,but it was also found to be unsatisfactory.During the 1993 session, a radically revisedfinance system was enacted. It provides foran “optimum” funding level and requires all

Of the statessurveyed,

Kentucky ranksfirst on

reduction of

disparities andamong thehighest in

increases in perpupil

expenditure.


Mary Fulton:...“courts now are

taking a deepinterest in theproblems with

and solutions toschoolfinance....Courts

are sendinglegislators back tothe drawing

table...”

declines in disparity will occur in the nextthree years. (Data do not show this, prob-ably because the effects of the suit are morerecent than the data used.)

TEXAS: Texas has seen tremendousequity gains in the past five years. Most ofthe gains have been achieved through“leveling down.” The court decisions of1989, 1991, and 1992 did not mandate anincrease in state funding. Every year sincethe Edgewood v. Kirby decision, there hasbeen an increase in school funding, but thesource of this has come largely from localrevenues. The school finance formulasreward increased tax effort and increasedthe required local share tax rate. The localrevenues are shared with other schooldistricts in substitution for state aid.

WEST VIRGINIA: The decision in the1980s has been “hailed as a revolutionarystep toward ending funding disparities.”(Education Week, Feb. 15, 1995) Unlikesome of the earlier cases in other states, thecourt was very specific about the task of thelegislature. As the preceding data indicate,great strides were taken to provide anequitable and adequate education, but thelegislature was unable to sustain thesestrong beginnings. Therefore, new compli-ance legislation was filed in 1994.

SUMMARY: These gleanings from themedia and from telephone interviews withknowledgeable persons in the states bear outthe observations of those who have fol-lowed activities in the state legislatures.Mary Fulton (1994), Policy Analyst of theEducation Commission of the States, hasstated that “Court play a bigger role infinance.” In an article of that title, she saysthat in more recent cases, “courts now aretaking a deep interest in the problems withand solutions to school finance...In addition,many court decisions are presenting stateleaders with ‘recommendations’ for rewrit-ing funding systems...Courts are sendinglegislators back to the drawing table and

plaintiffs are refiling lawsuits until accept-able solutions are presented.”

Some Unanswered Questions

Every empirically-based piece ofresearch raises more questions than itanswers and this is no exception. It wasknown before beginning this effort that somestates (California) had reduced their dispari-ties between school districts dramatically, inrecent years. It was also known that somestates (Illinois) had experienced greatincreases in those disparities. In this andrelated research, it has been established, atleast to some degree of satisfaction, that theconstitutional litigation is one factor inexplaining these differences between states.No claim is put forth, however, that theconstitutional litigation is the only factoroperating to distinguish betweens states withregard to educational disparities. With thenotable exception of the Moskowitz study,there is simply not enough longitudinalresearch on disparities within states to drawfirm conclusions.

Logically, one knows that economicdevelopment within states is not uniform;some states have experienced much moreunequal economic development than haveother states. This is a complicated subjectthat takes one rather far outside the world ofschool finance and into urban and regionaleconomics. If educational researchers wishto follow the equity river down to its verysources, that is the direction in which theywill have to head their expedition.

Neither is it known to what extent largemicro-forces operating throughout the entirecountry have influenced these long termtrends in disparities between school districts.A seminal study by the Twentieth CenturyFund, Edward Wolf’s Top Heavy: A Study ofIncreasing Inequality of Wealth in America(1995), reveals the disturbing fact thatwealth concentrations are now nearly twiceas great in the United States as they are in


present, l7 states have successfully defendedtheir statutes against constitutional chal-lenges and more will likely do so in thefuture. Traditions of court deference tolegislative bodies in this area are quitestrong in a number of states. What must beconsidered especially carefully is the effectof an adverse ruling. While evidence ofbeneficial effects of a ruling for the plain-tiffs is extant, the effects of a strong rulingfor the defendants is not so well known.

Reading recent briefs of attorneysgeneral from around the country reveals astrong pattern in defense strategy that alsoneeds to be taken into consideration byplaintiffs. If case precedent is poor forplaintiffs, and it is in many states, it isalmost a certainty that plaintiffs will face amotion to dismiss based upon prior caseswhich give great presumption of legitimacyto the legislative bodies in these areas ofpublic policy. This will mean a longstruggle all the way up the ladder of judicialreview to even get into court with theevidence. A trial on the merits—on thefacts, that is—will often have to be sorelywon in the courts since this is not given as amatter of right. Even if a trial on the factstakes place, it is almost a certainty thatdefense will argue that the current founda-tion level provides a minimally adequateeducation; therefore, any amount of dispar-ity above the foundation level is irrelevant.The state will argue that, if the poorest-funded school district in the state is ad-equately funded, then no one’s constitu-tional rights have been harmed. It may evenbe admitted, very grudgingly by the state,that education just might be a fundamentalright under the constitution, but it is a rightonly to enter the playing field, it does notconstitute a right to a level playing field.

At best, the constitution may require thestate to provide an education which isadequate to meet the needs of basic partici-pation in the representative governmentalprocesses of the state. The individual has a

Great Britain—exactly reversing a situationthat held before World War II. Wolf’s studysupports other evidence of income inequali-ties that have grown rapidly in the UnitedStates since roughly 1976. What the connec-tions are between these large national trendstoward greater income and wealth inequalityand school finance is unknown. Very likelythis relationship will remain unknown aslong as school finance research continues tobe as parochial as it has been in the past.Hopefully, the Twentieth Century Fund orsome other private foundation will becomesufficiently interested to explore thesematters. It does not appear likely at themoment that the federal government will bewilling to follow this research lead.

Conclusion and PolicyImplications

Is it worth challenging the system offunding K-12 education through the courts?On purely empirical grounds, the answer isprobably yes. The weight of the evidenceprovided in this study certainly suggests thatprogress can be made in reducing expendi-ture disparities between school districts bycourt action. Much less certain is theevidence that increases in funding for allschool districts can be improved; that is, thatadequacy as well as equity is served by thistype of litigation. In some states, Kentuckyand Tennessee come immediately to mind,progress can perhaps be made on both goals,equity and adequacy, by the litigation route.In other states, only one or the other of thegoals may be served.

However, before rushing to the courthouse with constitutional complaint hotly inhand, there are a number of very soberingconsiderations to keep in mind. In the firstplace, this legal process can take a very longtime and can be quite expensive. It is not atall unusual for this litigation to last a fulldecade and take well over a million dollarsin legal fees in any given state. Further, thechances of losing the case are quite high. At

At present, 17states have

successfullydefended theirstatutes against

constitutionalchallenges...


right that is only to the minimum educationnecessary to be a good citizen. Plaintiff thenfaces the steep challenge of showing thateducation at the foundation level is notadequate for voting, freedom of speech, etc.It is very difficult in a court of law todetermine just how much needs to be spentto arrive at minimally prepared citizens whoare capable of exerting their civil rights.

So we ask again, is it worth fighting allthose years and spending all that money?On ideological grounds rather than empiricalgrounds, the answer is a strong “Yes.” In thefirst place, this matter of inequalities be-tween school districts is an old, old woundthat lies festering in the body politic of theUnited States and must some day be excised.In 1823, when Thomas Jefferson was onlythree years away from death, he described asa failure his own “Bill for the GreaterDiffusion of Knowledge,” which he hadgotten through the Virginia legislature inl777.2 It failed said Jefferson because theVirginia legislature would not tax at the statelevel to support the system, but insisted onraising the funds for public education at thelocal level. That would not work saidJefferson, “because the rich will not pay forthe education of the poor.” Unfortunately,not much has changed since Mr. Jefferson’sday.

Another reason to persist in addressingthe problem is because, quite frankly, it isdangerous to defer abjectly to legislativebodies in matters as central to the health ofthe Republic as is education. Thomas Paine(1995) reminded us of this, in the early daysof the Republic, when he noted that, whiletyranny of the executive branch was morelikely than tyranny of the legislative branch,the latter was not an impossibility. That wasespecially true, said Paine, when the legisla-ture had been in the hands of a single party

[the ”doctrine ofunique

function”] arguesthat publiceducation is like

absolutely noother publicservice. It is the

only publicfunction that hasits own article in

every single stateconstitution inthis nation.

2 The Virginia legislature did not get around to final consideration of Thomas Jefferson’s full educational proposals until1817. On February 20th of that year, most of his design for public education in Virginia was defeated. The agingJefferson expressed his disappointment both in his autobiography in 1821, and in a number of personal letters; forexample, his letter to General James Breckinridge in that same year. See The Writings of Thomas Jefferson, 1984, TheLibrary of America, NY, NY.

for some time. The fact of the matter is that,every once in a while, we need to be force-fully reminded that in the United States wedo not have a British parliamentary system.The founding fathers intended that all threebranches of government (executive, judicial,and legislative) be used for the solution ofpublic policy problems. These cases are inthe state courts because a significant part ofthe citizenry believes that their concernshave not been satisfactorily addressed by thevarious state legislatures. Many observersfeel the “gridlock” on school finance mattersis increasing not decreasing; therefore, theappeal to the courts is likely to continue.

Finally, these cases are needed as ameans to re-examine and possibly reassertthe “doctrine of unique function.” Attrib-uted to the late Charles H. Judd (1934) of theUniversity of Chicago, that doctrine arguesthat public education is like absolutely noother public service. It is the only publicfunction that has its own article in everysingle state constitution in this nation. Thefounding fathers of every state apparentlybelieved that there was something veryspecial about education that deserved thiskind of unique legal treatment in the states’constitutions. For many decades, theseeducation articles lay almost unlitigated,really almost forgotten. The educationarticles were occasionally used to justifytaxation to support public education, but forlittle else. Now, citizens are engaged inasking the several supreme courts of the 50states just what these education articlesreally mean. Why were they put there in thefirst place? Are they simply vague goals, ordo they constitute real mandates upon thestate governments to do something? Do theyconstitute fundamental rights and, if so, howdo they relate to other rights in the constitu-tion?


...it may benecessary to...

expand spendingon education...torestrict spending

in othergovernmental

functions.

It is also true that many of the plaintiffssincerely believe that education is not just“a” fundamental right among a number ofothers, rather they believe that it is “the”fundamental civil right. That is, it is notpossible to have a meaningful right to vote,to free speech, to freedom of the press, etc.without an adequate education. For theseindividuals education goes to the very heartof the relationship between an individual andhis or her government and, thus, qualifies asa fundamental right. Granted, it is mostlythose on the liberal side of the aisle whopress that particular line of argument, butconservatives are not at all left out in thisdoctrine. For conservatives, education is thesource of an individual’s ability to restrictthe size and effect of government, because,

without an adequate education, the indi-vidual citizen cannot effectively defend hisor her liberties against the incursions of thestate. Ultimately, the conservative sideswith Jefferson who said, “a nation thatexpects to be ignorant and free asks whatnever was, and never will be.” Paradoxi-cally, it may be necessary to actually expandspending on education in order to restrictspending in other governmental functions.“Mr. Republican,” the late Senator RobertTaft, used that argument in defending manyof his pro-education spending votes in theU.S. Senate.

Educators often view the world througha very narrow lens. This constitutionallitigation will continue because it swells upfrom the magma of this society, from publicpolicy problems never solved from the veryopening days of this Republic. Viewed inthis light, the constitutional challenges areserving a very useful function. Just likevolcanoes, they relieve the pressure beneaththe surface that would otherwise tear theRepublic apart. We should be thankful forthem.


References

Fulton, Mary L. Fall 1994. “Courts Play Bigger Role in Finance.” State Education Leader 13(2). EducationCommission of the States.

Heise, Michael. Fall, 1995. “The Effect of Constitutional Litigation on Education Finance: More PreliminaryAnalyses and Modeling” Journal of Education Finance 21(2): 195-216.

Hertert, Linda, Carolyn Bush, and Allan Odden. Winter 1994. “School Finance Inequities Among the States: TheProblem from a National Perspective.” Journal of Education Finance.

Hickrod, G.A., Edward R. Hines, Gregory P. Anthony, and John A. Dively. Fall 1992. “The Effect of ConstitutionalLitigation on Educational Finance: A Preliminary Analysis.” Journal of Education Finance 18(2).

Hickrod, G.A., R. Arnold, R. Chaudhari, J. Meng, and G.B. Pruyne. December 1994. A Firebell Ringing in theNight: Increasing Inequalities in Illinois School Finance. Center for the Study of Educational Finance. Illinois StateUniversity. Normal, IL.

Jefferson, Thomas. 1984. The Writings of Thomas Jefferson. The Library of America. New York: NY.

Judd, Charles. 1934. Education and Social Process. Harcourt Brace.

Paine, Thomas. 1995. “On the Affairs of Pennsylvania, 1786.” The Collected Writings of Thomas Paine. TheLibrary of America, New York: NY.

Peternick, Lauri. 1995. “The Effect of Overturning Educational Funding Practices Through the State Courts onOverall Per Pupil Expenditures.” Paper prepared for the 1995 annual meeting of the American Education FinanceAssociation in Savannah, Georgia.

Riddle, Wayne and Liane White. Winter 1994. “Variations in Expenditures Per Pupil Within the States: Evidencefrom Census Data for 1989-1990.” Journal of Education Finance.

Ward, James Gordon. 1990. “Implementation and Monitoring of Judicial Mandates: An Interpretative Analysis,”The Impacts of Litigation and Legislation on Public School Finance. Ed. J.K. Underwood and D.A. Verstegen.Harper and Row.

Wyckoff, James H. 1992. “The Intrastate Equality of Public Primary and Secondary Education Resources in theU.S., 1980-87.” Economics of Education Review 2(1).


APPENDIX A

STATUS OF SCHOOL FINANCE CONSTITUTIONAL LITIGATIONCompiled by G. Alan Hickrod, Robert Lenz, and Paul Minorini

June 1995I. Plaintiffs won at the state supreme court level (8 states):

Washington Seattle v. Washington, 1978Kentucky Rose v. The Council, 1989Connecticut Horton v. Meskill, 1977;

Sheff v. O’Neill, 1995Tennessee Tennessee Small School Systems v. McWherter, 1993, 1995Massachusetts McDuffy v. Secretary of Education, 1993Arizona Roosevelt Elem. School Dist. 66 v. Bishop, 1994Texas Edgewood v. Kirby, 1989, 1990, 1991, 1992, 19953

New Hampshire4 Claremont, New Hampshire v. Gregg, 1991

II. Plaintiffs won at the state supreme court level, but further compliance litigationwas filed (6 states):

California Serrano v. Priest, 1971, 1977Wyoming Washakie v. Hershler, 1980West Virginia Pauley v. Kelly, 1979; 1988

Pauley v. Gainer, 1994Montana Helena School District v. Montana, 1989, 1993

Montana Rural Ed Assoc. v. Montana, 1993New Jersey Robinson v. Cahill, 1973;

Abbott v. Burke, 1985, 1990, 1994Arkansas Dupree v. Alma School District, 1983;

Lake View v. Arkansas, 1994

III. Plaintiffs lost at the supreme court level and there have been no further complaintfiled or lost (11 states):

Michigan Milliken v. Green, 1973 East Jackson Public School v. State, 1984

Idaho Thompson v. Engelking, 1975; Frazier et al. v. Idaho, 1990

Georgia McDaniels v. Thomas, 1981Colorado Lujan v. State Board of Education, 1982Wisconsin Kukor v. Grover, 1989Oregon Olsen v. Oregon, 1979;

Coalition for Ed. Equity v. Oregon, 1991Minnesota Skeen v. Minnesota, 1993

3 System found constitutional on latest supreme court decision.4 Win for plaintiffs at appeals on motion to dismiss.


North Dakota Bismark Public Schools v. North Dakota, 19935

Nebraska Gould v. Orr, 1993Virginia Alleghany Highlands v. Virginia, 1991 (Withdrawn 1991)

Scott v. Virginia, 1994Maine M.S.A.D.#1 v. Leo Martin, 1992, 1995

IV. Plaintiffs lost at the supreme court level, but there have been further complaints filed (6 states):

Pennsylvania6 Dansen v. Casey, 1979; 1987 Pennsylvania Association of Rural and Small Schools v.Casey, 1991;

Ohio7 Board of Education v. Walter, 1979 Howard v. Walter, 1991 Thompson v. State of Ohio, 1991 DeRolph v. State, 1992.

New York8 Board of Education v. Nyquist, 1982; 1987 Reform Educational Financing Inequities Today (R.E.F.I.T.)v. Cuomo, 1991 Campaign for Fiscal Equity v. State of New York, 1993

Maryland Hornbeck v. Somerset County, 1983 Bradford v. Maryland State Board of Education, 1994

South Carolina Richland v. Campbell, 1988 Lee County v. Carolina, 1993

North Carolina9 Britt v. State Board, 1987 Leandro v. State, 1994

V. Litigation was present, but no supreme court decision has been rendered (10 states):

Illinois6 The Committee v. Edgar, 1990Alabama7 Alabama Coalition for Equity v. Hunt, 1990;

Harper v. Hunt, 1991Alaska Matanuska-Susitna Borough v. Alaska, 1989South Dakota10 Bezdichek v. South Dakota, 1991Rhode Island7 City of Pawtucket v. Sundlun, 1992Missouri The Committee v. Missouri and

Lee’s Summit P.S.U. v. Missouri, 199411

Louisiana Charlet v. Legislature of State of Louisiana, 1992Florida8 Coalition v. Childs, 1995New Mexico Alamagordo v. Morgan, 1995Vermont Lamoille County v. State of Vermont, 1995

5 Majority (3) ruled in favor of plaintiff, but North Dakota requires four justices to declare a statutory law unconstitu-tional.

6 Win for defendants at appeals on motion to dismiss.7 Win for plaintiffs at district on merits.8 Win for defendants at district on motion to dismiss.9 Win for plaintiffs at district on motion to dismiss.10 Win for defendants at district on merits.11 After a trial on the merits, the trial court rendered a decision for the plaintiffs, but reserved many issues for a later

hearing. The defendants appealed the trial court’s decision, and on June 21, 1994, the Missouri Supreme Courtdismissed that appeal on the grounds that the judgement below was not final.


VI. No litigation is present or case is dormant (9 states):

Delaware MississippiHawaii NevadaIowa UtahIndiana Lake Central v. Indiana, 1987 (Withdrawn)Oklahoma Fair School v. State, 1987Kansas Consolidated:

Unified School District 229, et al. v. State, 1992Unified School Dist 244, Coffey County, et al.v. StateUnified School District 217, Rolla, et al. v State

Category A: States in which the state supreme court has declared that educationIs a fundamental constitutional right (13 states):

Arizona Shofstall v. Hollins, l973Wisconsin Busse v. Smith, l976California Serrano v. Priest, l977Connecticut Horton v. Meskill, l977Washington Seattle v. Washington, 1978Wyoming Washakie v. Hershler, l980West Virginia Pauley v. Bailey, l984Montana Helena v. State, l989Kentucky Rose v. the Council, l989Minnesota Skeen v. Minnesota, 1993Massachusetts McDuffy v. Secretary of Education, 1993Tennessee Tennessee Small School Systems v. McWherter, 1993Virginia Scott v. Virginia, 1994

Category B: States in which the state supreme court has declared that education isNOT a fundamental constitutional right (10 states):

New Jersey Robinson v. Cahill, l973Michigan Milliken v. Green, l973Idaho Thompson v. Engelking, l975Oregon Olsen v. State, l976Pennsylvania Dansen v. Casey, l979Ohio Board v. Walter, l979New York Levittown v. Nyquist, l982Colorado Lujan v. Colorado, l982Georgia McDaniel v. Thomas, l982Arkansas Dupree v. Alma, l983


Category C: Lower court decision on education as a fundamental right:

1. States in which a circuit or appellate court has declared that educationIS a fundamental right (7 states):

Alabama Alabama Coalition for Equity v. Hunt, 1990Missouri Committee v. Missouri, 1993Minnesota Skeen v. Minnesota, 1992North Dakota Bismark Public Schools v. North Dakota, 1993Washington Tronsen v. State of Washington, 1991Ohio DeRolph v. State, 1992.Rhode Island City of Pawtucket v. Sundlun, 1992

2. States in which a circuit or appellate court has declared that education IS NOT afundamental right (2 states):

Illinois Committee v. Edgar, 1992New Hampshire Claremont, New Hampshire v. Gregg, 1991

Student-Level SchoolResource Measures

Selected

Papers in

School

Finance

62Selected Papers in School Finance, 1995

Robert Berne is Vice President forAcademic Development at New YorkUniversity, where he has been a facultymember since 1976.

He is the co-author of The Measurementof Equity in School Finance (Johns HopkinsUniversity Press, 1984) and the co-editor ofOutcome Equity in Education (CorwinPress, 1994). In the field of public sectorfinancial management, he has co-authoredThe Financial Analysis of Governments(Prentice-Hall, 1986) and authored TheRelationships between Financial Reportingand the Measurement of Financial Condi-tion (Governmental Accounting StandardsBoard, 1992). In addition, he has publishedwork in numerous journals and testified inschool finance court cases.

Dr. Berne served as executive director ofthe New York State Temporary Commissionon New York City School Governance, alsoknown as the Marchi Commission, from1989 to 1991 and was the director of policyresearch for New York State’s TemporaryCommission on the Distribution of SchoolAid (Salerno Commission) in 1988.

Dr. Berne helped to develop, and is thechairman of, the Steering Committee of theSoros Foundation-funded Institute for LocalGovernment and Public Service. TheInstitute, located in Budapest, Hungary,supports the fields of local government andgovernance by assisting organizations inCentral and Eastern Europe.

of school-level allocations and budgeting,and public finance issues in K–12 education.She is author of Statistical Analysis forPublic and Nonprofit Managers (Praeger,1990), co-author of The Measurement ofEquity in School Finance (Johns HopkinsUniversity Press, 1984) and writes fre-quently for journals such as the Journal ofEducation Finance, Educational Evaluationand Policy Analysis, and Journal of PolicyAnalysis and Management.


Robert Berne and Leanna StiefelRobert F. Wagner Graduate School of Public Service

New York University

Leanna Stiefel is professor of econom-ics at the Robert F. Wagner Graduate Schoolof Public Service at New York Universitywhere she directs the public and nonprofitprogram for M.P.A. students. Her areas ofspecialty include education finance, publicfinance and financial management, andapplied statistics.

In her education finance research, shefocuses on the equity of state and intra-district level finance systems, the efficiency

About the Authors

Student-Level School Resource Measures 63

The purpose of

this paper is todiscuss the typesof student-level

resourcemeasures thatwould be

preferred ifstudent-levelresource data are

collected on aregular basis.


Robert Berne and Leanna StiefelRobert F. Wagner Graduate School of Public ServiceNew York University

experience of teachers, but seldom is there arelatively complete account of all therelevant resources received by an individualstudent.

The purpose of this paper is to discussthe types of student-level resource measuresthat would be preferred if student-levelresource data are collected on a regularbasis. In section one, the paper explores thetypes of questions that could be answeredwith student resource data. Section tworeviews selective literature to show howanalysts have approached answers to suchquestions with data now available. Sectionthree discusses cost accounting concepts thatare useful for thinking about how to collectappropriate student-level resource data.Section four recommends alternatives forNCES to consider. These alternatives takeinto account other related NCES datacollection efforts, such as the NationalEducation Longitudinal Study of 1988(NELS:88), National Assessment of Educa-tional Progress (NAEP), and Common Core

Americans are obsessed with theirpublic schools. They are especially focusedon productivity—what student outcomes dotheir tax dollars buy? How can betterstudent performance be obtained with thesame or lower spending? If more money isto be spent, where will it do the most good?In response (or anticipation) to the public,analysts are studying the question of whichresources and types of organizations aremore productive than others, but a majorproblem with these studies is the unavail-ability of appropriate resource data. Eitherthe data are collected for another purposeresulting in an incorrect unit of analysis, orthe data lack sufficient detail. Data are mostcommonly available at the district level on aper pupil basis. However, these data areaveraged over a great variety of studentsand programs and, as a result, they providevery imprecise measures of resourcesconsumed by any particular student in adistrict. In addition, student-level data areoften incomplete. There may be measuresof average total expenditures or average

Introduction


consuming? Both attribution andconsumption questions are important.A resource may be attributed (that isassigned or allocated to a student) butthe student may or may not actuallyconsume (receive) all of that resource.The consumption question is even moredifficult to answer than the attributionquestion. A key issue is whetherstudent differences and resource differ-ences can be matched to obtain validand reliable vertical equity and equalopportunity measures.

3.Resource intent questions: Forstudents who have been identified asentitled to specific resources (e.g.,handicapped, bilingual, or compensa-tory education), do they receive addi-tional resources? Does this match theintent of the financing system? Thisquestion can be addressed withinschools and districts and across stu-dents, schools and districts.

These three types of questions allrequire some kinds of student-level resourcemeasures, that is, resource measures thatcapture the variation across differentstudents. All of these questions have beenaddressed in previous studies, not alwayswith the appropriate data. New data collec-tion efforts should attempt to address theconcerns raised by each of the questions.

Selective Literature Review ofStudies that Use Student-LevelResource Allocation Measures

The purpose of this section is to identifythe kinds of resource data that have beenused in previous production, equity, andintent studies, and to evaluate whether thosedata have been adequate for the research. Afew well-known studies have been selectedin each area and, therefore, this section isillustrative rather than comprehensive. Wewish to both identify the need for student-level data and show how analysts have dealt

of Data (CCD). Section five concludes thepaper.

How Would Student ResourceMeasures Be Used: QuestionsStudent Resource Data CouldAnswer

There are three types of frequently askedquestions whose answers require student-level resource data. The three questionsconcern production functions, equity ofresource distribution, and intent of resourcedistribution. All three are asked by policymakers, the public, and analysts. Thissection describes the nature of the questions,and the next section reviews selectiveliterature where they have been addressed.

1.Production function questions, withboth the student and school as the unitof analysis: What is the relationshipbetween outputs and inputs, especiallyschool inputs? Several more specificquestions require production functionknowledge:

•Resource effectiveness questions:Examples: Do additional resources forchildren lead to additional outcomes?What kinds of resources and resourceuse lead to the largest outcome gains?These questions can be at a generallevel, or for a specific program (profes-sional development or mixed gradeclasses), or for a specific type of student(at-risk or bilingual).

•Cost-effectiveness questions: What isthe cost effectiveness of one programversus another, for example, readingrecovery versus cross-age tutoring?

2.Equity questions: Examples: Withina district, for students in differentschools, or for different students in thesame school, what are the resourcesdirectly attributable to them and are theyequitable? What resources are they

There are threetypes of

frequently asked

questions whoseanswers require

student-level

resourcedata...[these]

concern

productionfunctions, equity

of resource

distribution, andintent ofresource

distribution.


with that need in the relatively recent past.

Production Function Studies

Table 1 summarizes the types of re-source variables identified in the studies wereviewed. The categories of variables inTable 1 are taken from a recent study byRichard King and Bettye MacPhail-Wilcox(King and MacPhail-Wilcox 1994). The restof this section gives a brief description ofeach article and our conclusions aboutresource variables based on the article.

One of the most “famous” recentproduction function studies is EricHanushek’s 1986 review of 147 estimatedequations in the Journal of EconomicLiterature (Hanushek 1986). From the pointof view of types of resource data used,Hanushek found that most studies includedthree school inputs that are the largestdeterminants of instructional expenditures—teacher experience, teacher education, andclass size. In addition, approximately 55percent of the equations he reviewed in-cluded either teacher salary or expendituresper pupil. Because of additional evidencethat some characteristics of teachers aresignificant determinants of changes instudent achievement, Hanushek discussesstudies that include teacher specific dummyvariables as a way of getting at a “totalteacher effect.” For the purposes of thispaper, it is important to note that most of thestudies reviewed use data that are availablein administrative records for districts orschools. Presumably more detailed data atthe student level would be beneficialbecause they would allow researchers toexplore whether individual components ofschool resources affect outcomes. At aminimum, some detail on teachers should bemade available.

King and MacPhail-Wilcox (1994)recently reviewed the results of a largenumber of production function studies.They describe a wide variety of school

inputs that researchers have used. Many ofthese are resource variables, or closelyrelated to such variables. The variables aresummarized in Table 1, with the identifiersK and M. For teacher characteristics, Kingand MacPhail-Wilcox note that years ofexperience, training, verbal achievement,and salary are the most often analyzedvariables and that “teacher experience andteacher abilities bear a stronger, and moreconsistent, relationship with pupil perfor-mance on achievement tests than do othercharacteristics.” (p. 53) For policy andadministrative arrangements, the authorsnote that class size, pupil-teacher ratio, andability grouping have been analyzed fre-quently. For classroom-based research, theauthors write:

Researchers have argued for manyyears that studies would beimproved if individual childrenand classrooms were the unit ofobservation rather than the schoolor district,... if resources wereidentified as those available to aspecific child rather than byaverage resources in a school ordistrict, and if processes were toinclude the quality and intensityof student-teacher interactions andtime on task. (p. 59)

From this study, we conclude thatresearchers from different disciplines haveemphasized different kinds of resourcevariables. Purely financial resource vari-ables seem to have been taken from availablerecords; more detailed classroom data areobtained by special, mostly one-shot,studies. Once again, it would be valuablefor one source to combine as many of thesedifferent kinds of variables as possible andto combine them in an on-going way.

In an article that concludes “that whentargeted and managed wisely, increasedfunding can improve the quality of publiceducation,” Ronald Ferguson (1991, 488)

...three schoolinputs...are thelargest

determinants ofinstructionalexpenditures—

teacherexperience,teacher

education, andclass size.


uses district-level data from Texas toestimate production function relationships.The resource variables he uses are summa-rized in Table 1, and are indicated by F.Regarding data availability, Ferguson beginshis article:

Allocating resources efficientlyand equitably in public primaryand secondary schools has beenan elusive goal. Among theprimary reasons is the surprisingscarcity of data appropriate forestablishing the relative impor-tance of various schooling inputs.As a result, recent research todiscover how increasing spendingmight affect the quality of school-ing and how improving the qualityof schooling might affect howmuch children learn has reana-lyzed old data or has relied ondata sets that are very limited insize and scope. (p. 463)

Ferguson makes a case for his district-level data in Texas, but also notes the needfor disaggregated data in some cases. Weconclude that even when researchers arecareful to put together a comprehensive andunique data set, they cannot always obtainresource variables at the correct unit ofanalysis.

In an innovative study, Byron Brownand Daniel Saks use detailed time alloca-tions to students and subjects they study andtheir relationship to learning in reading andmathematics (Brown and Saks 1987). Theauthors explore technology and teachervalues within the classroom. The resourcedata they use are summarized under Class-room-Based Research in Table 1, and areindicated by B and S. The study addressesthe issues in this paper because it is one ofthe few that collects details at the studentlevel and because it finds small but signifi-cant positive effects on learning when more

...even whenresearchers arecareful to put

together acomprehensiveand unique data

set, they cannotalways obtainresource variables

at the correct unitof analysis.

Table 1.–Resource Variables Cited in Recent ProductionFunction Studies

Variable¹Study²

Characteristics of Teachers TeacherexperienceH, K-M, F Teacher educationH, K-M, FTeacher salaryH, K-M Total teacher (dummy)H, FVerbal achievementK-M PersonalityK-M Employ-ment statusK-M Turnover rateK-M SocioeconomicstatusK-M Job satisfactionK-M NTE or other testscoreK-M, F Financial Measures and FacilityCharacteristics Expenditures per pupilH, FAdministrationF Instructional serviceF Cur-riculum activitiesF TransportationFMaintenanceF Quality of facilitiesK-M Policy andAdministrative Arrangements Class sizeH, K-M, FPupil-teacher ratioH, K-M, F Ability groupingK-MClassroom-based Research Curriculum contentK-M Instructional groupingK-M Learning materialsK-M TimeK-M Teacher expectationsK-MHomeworkK-M Reinforcement practicesK-MPerformance feedbackK-M Proportion time studentsengagedB-S Proportion time self-pacedB-SProportion time substantiveB-S Proportion timestudents groupedB-S Proportion time studentstutoredB-S Minutes per subjectB-S ProcessvariablesM

¹The classification of variables is taken from King

and MacPhail-Wilcox, 1994. ²H = Hanushek, 1986;K-M = King and MacPhail- Wilcox,1994; F = Ferguson, 1991; B-S = Brown and Saks,

1987; M = Monk, 1992. SOURCE: Robert Berne and Leanna Stiefel,unpublished tabulations.


time is spent on a subject. Thus, there mayindeed be a need to collect such detailed datamore systematically and over time, so thatothers can explore their effect.

Brown and Saks make the following(now familiar) comments on the type of dataavailable for production function studies:

But even the best of the previousstudies have suffered from one ormore serious defects. First, theproduction functions have relatedoutput ... not to the inputs ofteaching actually received by thestudent but to some average inputavailable to the class or school.Because the allocation of inputswithin schools and classrooms isitself highly variable, such datacan provide highly misleadingestimates of a school’s underlyingproduction technology...Second, previous work has beenseverely limited in its ability todeal with the consequence ofheterogeneity of teachers andstudents....Third, because theteacher is producing multipleoutputs in the classroom... Itbecomes fundamental... to dis-cover what the teacher is trying toaccomplish in that classroom, todiscover just what the teacher’sobjectives or preferences are. (p.320)

Both through what they declare andwhat they find through the use of their data,Brown and Saks make a strong case forcollecting data at a very micro level.

Finally, a recent article by David Monkexamines the history (to the present) ofproduction function studies in the context ofthe national interest in productive schools(Monk 1992). He notes that studies ofprocess, often at the classroom level, seem to

be less noticed than the more commonproduction function studies that use moreaggregated data. He also notes a recenttrend back toward more aggregate data inpublished articles about production func-tions. Finally, based on his assessment ofthe history of production function research,he advocates a classroom approach, whichwould involve collection of student andclassroom process data.

The pattern of inconsistent andlargely insignificant resultsreported in this article points in apromising direction for futureproductivity research in educa-tion, and this direction involvesraising the classroom to a higherlevel of importance in the conductof productivity research. Thus, Iam calling for a more disaggre-gated approach than has beencharacteristic of recent attempts toestimate production functions. Iam also raising a concern overplacing too much emphasis onschool-level analyses, somethingthat I believe has happened as aby-product of early effectiveschools studies. And I am arguingthat more can be done with theeconomically oriented processstudies that I reviewed earlier.My goal here is to motivate aclassroom-oriented line of inquiryinto education production that isdeductively driven and thatcomplements the already devel-oped school-oriented studies.(p.320)

In conclusion, these authors, and theauthors whose studies they reviewed, seemto agree either explicitly or implicitly on thefollowing points. First, studies often useadministrative data because it is availablerather than because it is right. Second, datathat are disaggregated to the student level are

My goal here isto motivate aclassroom-

oriented line ofinquiry intoeducation

production thatis deductivelydriven and that

complements thealreadydeveloped school-

oriented studies.(Monk p.320)


better than data that are averaged from thedistrict level. Third, a wide range ofresources might be relevant to the determi-nation of gains in performance. The num-ber depends both on the exact nature of theproduction relationship that is hypothesizedand on empirical findings on what variablesare significant.

Equity Studies

There is a long history of school financeequity studies, most of which have useddistrict-level data. More recently, resourceequity across schools is being examined, forexample in court cases, such as the one inLos Angeles, or in school-based reforms,such as those in Chicago and Kentucky. Wereviewed two studies that have used school-level data to look at variations acrossschools and ask if the data have beenadequate to the task.

Berne and Stiefel (1994) analyzed thedistribution of New York City budgets andexpenditures per pupil by school and sub-district (community school district). Theyused published budget data and administra-tive data on expenditures that separatedstreams of funding into general education,special education, and reimbursable funding(e.g., Chapter I, state compensatory educa-tion etc.). They related the spending perpupil to the percent of pupils in poverty (byschool or subdistrict) to assess one aspect ofvertical equity.

Several problems with the data emerge.First, pupil counts for special education donot match funding streams. Second, al-though reimbursable funding can be attrib-uted to schools, there is no way to know ifthe funding is spent on the children forwhom it is targeted. Third, much of thedistrict spending is not allocated at theschool (or subdistrict) level. For example,fringe benefits, transportation, schoollunches, and utilities are not allocated. Astudy such as this could be done with much

more accuracy if there were student-levelresource measures that were defined to beinclusive and to differentiate between kindsof programs and students. The data wouldbe useful if it were gathered at the schoollevel or, if it were a sample of individualstudent-level data that was representative atthe school level. As part of this study ofNew York City school budgets, a model ofdisaggregated resource variables needed tobe developed. Whether or not NCEScollects such data for school districts, theexistence of a model way to representresource variables would help schooldistricts produce their own data and analy-sis.

Lawrence Picus, as part of a largerproject, studied school-level allocationsusing existing government data (the NCESSchools and Staffing Survey, 1987–88 andthe U.S. Census Bureau’s 1987 Census ofGovernments) (Picus 1993). Picus’ goalwas to look at patterns; he was not explicitlylooking at equity. We place the study in theequity category because some of the pat-terns can be reframed to reveal answersabout distributions that are generallythought about in terms of equity (distribu-tions with respect to demographics orgeography). The school is the unit ofanalysis in the Picus study and substantialeffort was required to construct the school-level data. The merged data set containsconsiderable resource information, but aswith the Berne and Stiefel study, more couldbe learned if there were a standard way ofreporting school-level resource data. WhilePicus is not explicitly interested in student-level data, it would improve his analysis,because the school-level data in the studyare averages over different kinds (andproportions) of students receiving differentkinds of resources (e.g., some schoolsreceive Chapter I funding and some do not).

We conclude from the school-basedstudies reviewed here that a well-defined setof student resource variables would improve

More recently,resource equity

across schools is

beingexamined...in

court cases...or in

school-basedreforms...


equity studies at the school level includingstudies that use administrative data, particu-larly if those variables are capable of servingas models for other data sets. They wouldbe useful if collected by NCES for equitystudies only if they were representative ofschools in a specific unit such as a district orstate, or inclusive of all schools in thoseunits.

Intent Studies

The four studies we review in thissection have used unique cost collectionmethods for answering questions about howresources flow to programs or schools. Onlyone study (Chambers et al. 1993) is designedto explicitly answer an intent question. Theother studies are classified as intent becausethey could be reframed to answer suchquestions.

The resource cost methodology (RCM)utilized by Jay Chambers and ThomasParrish and based on an ingredients ap-proach developed by Henry Levin (1983) isemployed to study the use of Chapter Ifunding (Chambers et al. 1993) and the costof alternative programs for students withlimited English proficiency (Parrish 1994).The method uses a bottom-up approach thatbegins with the program and client ofinterest and assigns costs to those programs.RCM is data intensive and generally expen-sive to employ. In the two studies reviewedhere the authors have used what they call a“purposive” rather than a random sample ofdistricts or schools, in part because of theneed to collect extensive data rather thanusing existing administrative records.Clearly this collection effort increases thecost of a study.

The RCM does provide one way to getaccurate cost data by program or client. Asdescribed by Parrish (1994):

Essentially, the resource costmodel system used in this studyinvolves three steps: 1) disaggre-gating and listing the relevant setof service delivery systems ormodels required for any educa-tional program, 2) determining thespecific resources utilized in eachdelivery system, and 3) attachingprices to each of these resourcesto determine specific programcosts. Overall and per pupil costsare determined on the basis ofthese programmatic standards andthe number of pupils enrolled ineach program. (p. 260)

If NCES were to develop a model wayof collecting resource costs by students thatincluded types of programs and types offunding for each student, researchers couldapproximate the RCM method for thestudents in the sample. However, the sampleof students would need to be representativeof the programs or funding sources (orchosen randomly with respect to these) inorder to make generalizations. The reviewof this method is useful primarily because itemphasizes the collection of detailed ingre-dients of resources by program and illus-trates the need to go further for some pur-poses than an aggregate average per pupilexpenditure number.

Bruce Cooper and colleagues havedeveloped a School-Site Allocation Model(SSAM) that is a cost accounting frameworkfor obtaining costs by function at the schoollevel (Cooper et al. 1994). They havedeveloped and tested the model in 10 schooldistricts across the United States. The modelis capable of generating costs by type ofschool (elementary, middle, high school) byfive functions for the school site and for thecentral office. The five functions areadministration, facilities and operations,staff support and development, pupil sup-port, and instruction. The article does not

If NCES were todevelop a modelway of collecting

resource costs bystudents that

included types of

programs andtypes of fundingfor each student,

researchers couldapproximate the

RCM method for

the students inthe sample.


give estimates of the cost of collecting thesedata, but does note that school districts arerequired to “use the same definitions ofcost items and place them into the correctfunctional categories—meaning that clarity,constancy, and accuracy are essential.” (p.71) Thus, like the Chamber and ParrishRCM, the data are most probably better thanwhen obtained from normal administrativerecords, but the cost is high. And again, ifNCES develops a prototype for studentresource data collection, it may be easier forschools and districts to follow.

Yet another recent effort to find costs atthe school level is the analysis by Coopers& Lybrand on behalf of the Mayor of NewYork City (Coopers & Lybrand 1994).Coopers & Lybrand label their model theSchool District Budget Model (SDBM). Itcrosswalks existing New York City Board ofEducation budget data into three classifica-tions: functions, school type, and programcategory. The functions are: instruction-schools, instructional support-schools,operations-schools, operations-central anddistricts, pass-throughs, and debt service.The school types are: pre-k, elementary,middle school, high school, and non-school.The program categories are: special educa-tion, bilingual/ESL, other categorical, andregular education. In contrast with theRCM and SSAM, the SDBM collects nonew data. Rather it finds ways torecategorize existing data. This is animportant distinction, because it probablymakes the SDBM less expensive per unit ofdata, but also less accurate. If the originaldata provide no clear indication of costsgenerated at the school versus the district orcentral level, then some assumptions need tobe made to place the costs in those areas.The other two models must also makeassumptions, but they are made earlier inthe data collection process when categoriesof costs are created and schools or programsare instructed how to put each type ofbudget item into a category.

All three cost models reviewed here areefforts to obtain data by school program, orfunction data that are more disaggregatedthan those available from normal adminis-trative records. All three are expensiveefforts and there is some possibility that acomprehensive school resource definitionby NCES could begin to lessen the expenseby establishing a more well-accepted model.

We should also note that the threesystems struggle with similar issues regard-ing allocation of overhead costs, centraloffice allocations to schools or programs,and the appropriate level of disaggregation.In part their categories are driven by thepurposes of the models. Cooper et al. are inpart interested in determining what they callthe “productivity” of various kinds ofexpenditures and, in particular, hypothesizethat additional expenditures at the class-room level will be productive. As Coopers& Lybrand state, “Development of theSDBM evolved from the belief that theprimary mission of schools is the directinstruction and support of students in theclassroom. All other functions exist tosupport this basic mission.” (p. 2). Cham-bers et al. and Parrish use their RCM totrace costs to programs or to find out howmuch supplemental spending there is onspecial programs. There is certainly no oneright way to collect disaggregated data;methods depend on the purpose of the datacollection.

Useful Cost Concepts

The cost accounting literature providesmethods for measuring costs that are usefulin decision making. In the context of thispaper, cost accounting is valuable because itemphasizes the need to conceptualize theuse of the cost data before the data arecollected. This section reviews some of thekey distinctions among types of costs foruse in our assessment of alternative waysfor NCES to construct student resource

...cost accountingis valuablebecause it

emphasizes theneed toconceptualize the

use of the costdata before thedata are

collected.


measures.

The distinctions described are:

•differences between departmental andproduct costing;

•the ways that product full costs can besubdivided (direct and indirect; variable andfixed);

•the ways that components of productfull costs can be allocated to students(process versus job-order costing andmethods in between);

•real resource versus dollar costs;

•the relevance of allocating componentsto students.

Departmental costing finds the costsof administrative units, such as agencies,departments, responsibility centers, etc. Aprimary purpose of departmental costing isto help managers administer units effi-ciently. In other words, decisions usingdepartmental cost data are centered on howto perform the department’s activities withan efficient use of resources. Productcosting, on the other hand, finds the costs ofproducing various kinds of products. Someprimary purposes of product costing are toappropriately price a product, to determinereimbursement levels, and to help decidewhether a product should be produced in-house, obtained by outside contract, or notproduced at all.

In our study, the concepts of productcosting are relevant because the questions tobe answered with the resource data arecentered, for the most part, on the student(product) and not on the administration ofthe districts or schools (units) that “producethe education” for the students. We areinterested in whether changing the wayresources are allocated to students willchange outcomes for the students, whether

resource allocations to students are equi-table, and whether students are receiving theresources that are intended for them. Allthree of these questions require ways toanalyze resources linked to students ratherthan ways to assess the effective use ofresources by particular organizational units.Naturally, the two questions of managementeffectiveness and student costs are related;but the distinction is useful when decidinghow to design a resource measurementsystem.

Products are assigned their full costs if,after adding together the costs of all productsin the organization, the organization’s totalcosts are determined. Full costs can besubdivided in numerous ways, but twoimportant distinctions are direct versusindirect and variable versus fixed. These aretwo different and not completely overlappingdistinctions. That is, direct costs are notalways variable and indirect are not alwaysfixed.

Direct costs are those that can beassigned uniquely to one product. They areclearly incurred as a result of that product’sproduction. Indirect costs, on the otherhand, are incurred as a result of productionof many products; they cannot be assigned,except by a rule, to just one product. Theyare often called overhead costs. Whenstudents are the product, instructionalsupplies such as textbooks, pencils, andwriting pads are clearly direct costs whilethe time of administrators such as thesuperintendent’s time is clearly an indirectcost.

Variable costs change as more units ofthe product are produced. For students,these are costs that increase when morestudents are added to a school. The amountof food for lunch is an example of a variablecost. Variable costs may occur in steps(semi-fixed), for example, when class sizereaches the point that an extra aide orteacher is needed. In such cases, groups of

...direct costs arenot always

variable and

indirect [costs]are not always

fixed.


additional students result in changes in costs.Fixed costs stay constant as more units of aproduct are produced. For students, theschool’s physical plant is a fixed cost (untilcapacity is reached). The cost of the princi-pal is a fixed cost. The distinction betweenfixed and variable costs depends on the timeframe and on the circumstances. Physicalplant becomes a variable cost at the pointwhen capacity is met or when the plantneeds replacement. The cost of the principalbecomes variable when more studentsrequire the hiring of assistants to helpmanage the school. Some costs have bothvariable and fixed components (semi-variable costs). For example, transportationcosts for students involve a fixed component(the buses and drivers) and a variablecomponent (the gasoline required to travel topick up additional students). Again, thefixed costs become variable when the bus isfull or needs replacing.

There are two systems for assigningdirect costs to products. In reality somecombination of the two systems is generallyused, but when conceptualizing a costingsystem, the distinction between the twosystems is useful. Job-order costingdetermines the costs of each individual unitof a product. A common example is thedetermination of the cost of hand-made,custom ordered furniture, where the specificmaterial, and amount and kind of laborwould be determined for each piece offurniture. Process costing determines thecosts of groups of identical units and thendivides by the number of units to obtain anaverage cost. A common example ofprocess costing is the determination of thecost of a box of a certain kind of breakfastcereal. The time needed to track costs toeach box would be great and since there isunlikely to be much difference betweenboxes, it is unnecessary for decision making.Job-order costing provides more accurateinformation; however, it is more difficult and

expensive to collect. Many actual costaccounting systems are hybrids of job orderand process costing.

Assigning resource costs to types ofstudents could be done either way, althoughsome process costing would probably beinvolved even when job-order system isdominant. For example, we could takeindividual students classified by type ofschool, type of special funding if any, typeof academic program, etc. and then use asample to determine as exactly as possiblehow much teacher time, other personneltime, and instructional supplies are used forthe student in a period of time. However,there might be some direct costs, for ex-ample physical education teacher time, forwhich an assessment of time spent witheach individual student would be unneces-sary. These costs might be assigned towhole groups (classes or schools) of chil-dren and then divided by the number ofchildren, producing process costing in thisdimension. At the other end of the spec-trum, process costing could be used atvarious levels of aggregation. For example,costs could be assigned by classroom (orgrade, school, district, academic program, ortype of funding) and then divided by thenumber of students.

Real resource costs are stated in termsof a resource’s natural units. Teacherresources would be determined as positionsor time per student; computers as numbersof computers or amount of time per student.The advantage of real resource costs is thatcomparisons across time or parts of thecountry can be made without worryingabout the different purchasing price ofdollars. For example, if beginning teachersof equivalent quality are paid differently insouthern versus northeastern states, costsstated as positions or time will more accu-rately reflect what students receive. On theother hand, dollar costs have advantages.

The advantage ofreal resourcecosts is that

comparisonsacross time or

parts of the

country can bemade without

worrying about

the differentpurchasing price

of dollars.


Dollars are a common measure; they allowus to aggregate different kinds of resourcesinto one number. In addition, they can beadjusted for differences in purchasing powerand if they are, they can reflect differencesin quality in ways that counting positions orminutes cannot. For example, a moreexpensive teacher might also be a moreeffective one—and dollars can help indicatethis.

Recommendations for NCESStudent-Level ResourceMeasures

This section builds on the analysis of theuses of resource measures, the selectiveliterature review, and the basic componentsof cost accounting systems to formulaterecommendations for NCES as it considerswhether to invest in the collection of stu-dent-level resource data. In addition to thereview of the policy literature in section two,we also have reviewed various NCESactivities that relate to the development of astudent-level resource measure. First,Chapter 5, “Cost Accounting for EducationalPrograms,” of Financial Accounting forLocal and State School Systems, 1990,presents an application of cost accounting toLEAs, and discusses uses, designs, andapplications. Second, NCES has a success-ful history of student-level data collections,most notably the National LongitudinalStudy of the High School Class of 1972(NLS-72), High School and Beyond(HS&B), and the National Education Longi-tudinal Study of 1988 (NELS:88). Bothactivities have implications for our recom-mendations.

Our recommendations should be viewedin a cost-benefit framework. While mindfulthat NCES has limited resources, a fullassessment of alternative courses of action isclearly beyond the scope of this paper. Theconclusion brings together our recommenda-tions, with some consideration of costs,although it is a partial view from the NCES

perspective.

The initial question that frames therecommendations on the development of astudent-level resource measure is the pur-pose to which it will be used. This isparticularly important here because thedesign of cost systems can vary according totheir purpose. We identified three broadpurposes in section two: effectiveness ofresource use, equity, and intent. All threepurposes could be served by a student-levelresource measure given the questions facingour education system, current state ofknowledge, availability of data, and potentialbenefits from a student-level resourcemeasure. However, the question of effectiveresource use should be placed at a slightlyhigher priority than the other two purposes.It may be that an effort to develop a student-level resource measure can serve all threepurposes simultaneously, but we suspect thatchoices will need to be made and need totake into account the higher priority purpose.

If the student-level resource measurewill be used to assess the effectiveness ofresource use, then the measurement systemwill need to link resources with studentoutcomes. That is, not only is it necessary tomeasure resources, usually in terms of dollarcosts, it is necessary to know the differentoutcomes that are achieved by students sothat the outcomes and costs can be linked.Note that NCES already has experience withcomplex data sets with the student as theunit of analysis, where outcome measuresare included (e.g., NELS:88).

Based on this highest priority purpose,the effectiveness of resource use, NCESshould move ahead to develop a frameworkfor a student-level resource measure, regard-less of whether the data are actually col-lected. This development should take placebefore a decision is made to actually collectthe data to permit substantial input into itsdefinition and design. Given the likelyamount of data that will need to be collected

The initialquestion thatframes the

recommendationson thedevelopment of a

student-levelresource measureis the purposes to

which it will beused.


and the expected cost of such an effort, thereshould be broad input into the developmentof this framework. Note that even if theactual data collection effort by NCES doesnot move ahead, the development of aframework could benefit other efforts at thestate or local levels to collect data onstudent-level resources.

A student-level resource measure willneed to build upon the existing financialaccounting system. The financial accountingsystem produces the basic elements that willbe categorized and combined into a costaccounting system. Thus, the student-levelresource measure will need to take intoaccount the basis of accounting (usuallymodified accrual), which means that thereare not precise distinctions between operat-ing and capital items, and the fund structure,which sometimes creates accounting-baseddivisions of expenditures that are not appro-priate for a student-level resource measure.

The framework for a student-levelresource measure will need to consider thefollowing elements. First, will the measurebe based on organizational units (departmen-tal costing) or students (product costing)?Given the primary purposes for gathering thecost information, a student-based system(product costing) will be required. Whilean organizational model can provide infor-mation on certain divisions of resources,such as instruction versus non-instruction,without more specific links to outcomes atthe student level, questions of resourceeffectiveness cannot be addressed at anappropriate level of disaggregation.

Second, will the student-based, student-level resource measure distinguish amongthe following concepts?

•Full versus partial costing. Todetermine resource use, it is desirable toobtain measures of full costs, not a partialmeasure that may exclude some “overhead”or other costs. The need for full costs is

important because when these costs arebroken down, the list will give analystschoices of resource definitions identicalacross students. If full costs are not gath-ered, then there is a risk that different typesof overhead costs will be omitted by differ-ent schools or that the specific list ananalyst would find valuable is missing. Afull cost definition will need to be appliedconsistently in all settings. In severalareas, (for example, teacher pensions, whichmay be partially or wholly state-financed),the state and local accounting systems maymake inclusion of full costs difficult.

•Direct versus indirect activities. Webelieve that this is an essential distinction toassess resource effectiveness, and to trans-late the results of the analysis into practicalrecommendations. Indirect costs must oftenbe allocated by formula to students and thuswill not always be useful to analysts. Also,with this distinction and the further delinea-tion of types of direct and indirect costs, theeffectiveness of different resources and indifferent combinations can be assessed.Any cost accounting system will be better(reliably and validly) able to specify directcosts compared to indirect costs. If thedistinction is used, the data should alsocontribute to the debates over whetherresources spent in the classroom, or oninstruction, or for specific types of instruc-tion (such as professional development), aremore productive than other patterns ofresource use. Note that the conceptualdistinction may be useful but may beaffected by the nature of the accountingsystem. For example, fringe benefits (healthcoverage, pensions, etc.) would be ideallydefined as direct, but they may be accountedfor in a manner (for example, at the state ordistrict level) such that they are precluded ina direct costing method.

•Different educational programs. Thiswould include, for example, regular instruc-tion at elementary, middle and high schools,special education at different levels, voca-

Thus, thestudent-level

resource measure

will need to takeinto account the

basis of

accounting...whichmeans that there

are not precise

distinctionsbetween

operating and

capital items, andthe fund

structure...


tional education, etc., as well as supportprograms such as school administration,district administration, facilities mainte-nance, transportation, etc. This distinctionis important from both educational (howlearning is organized) and organizational(how tasks are structured for management)points of view. All activities should becategorized in one (or more) programs.

•Process versus job order costing.Recall that process costing is used when theresource use does not vary across the unitsusing the resources and the units themselvesdo not vary. For most educational issues,both the resource use varies (differentialteacher time for different pupils) and theunits vary (students are different from oneanother). Thus, to the degree possible, astudent-level resource measure shouldemploy a job order costing system to reflectthe variation in resource use across differentpupils.

•Alternative funding sources (e.g.,local, state, and federal). From a policyperspective, this is an important variablebecause the revenue distribution systemsvary by level of government, but from a costaccounting point of view this may be one ofthe more difficult aspects of the resources tocapture. At the teaching and learning level,resources from all sources are mixed, and infact resources themselves cannot be exam-ined to determine the funding source in thesame way as they can for other distinctionswe examine (program, direct versus indi-rect, etc.). To some degree if we learn aboutresource effectiveness, it is a somewhatlower priority to learn the source of funding,but given the strong influence of the inter-governmental system, it would be desirableto include this distinction.

•Methods for allocating indirect costs.Indirect costs, by their nature require costallocation to obtain full costs. Key deci-sions in the indirect costing system beyondwhich costs are indirect include the basis for

the allocation (students, square feet, teach-ers, etc.), and the method for the allocation(one step, two step, etc.). The basis andmethod of allocation should be made veryclear; for many purposes analysts may wantto omit the indirect costs if the formula ismechanistic.

•Capital versus operating resources.One of the shortcomings of LEA accountingsystems, and all governmental accountingsystems in general, is the imprecise treat-ment of capital costs. Ideally, some annualcontribution of capital (for example theschool building or the school buses) shouldbe included in a student-level resourcemeasure. Unfortunately, because accountingconcepts such as depreciation are not useduniformly throughout school district ac-counting systems, capital items will probablyrequire special treatment. While it ispossible to think of using proxies for depre-ciation, such as annual debt service (princi-pal and interest payments), one would needto proceed very carefully. It is quite possiblecapital is financed quite differently acrossdistricts in terms of the percentage that isdebt-financed versus the percentage that ispurchased with current funds.

...a student-levelresource measureshould employ a

job order costingsystem to reflectthe variation in

resource useacross differentpupils.


Also, some states may help finance debtwithout accounting for the debt service flowthrough the district’s financial records. Anycost accounting systems must be very carefulto treat capital uniformly and to make surecapital purchases do not not cause or maskvariations in other resources.

•Dollars versus real resources. Insection three dollars and measures of realresources, for example, number of teachers,guidance counselors, etc. were compared.We concluded that due to the differentaspects of resources that they each capture, itwould be desirable to measure both dollarsand real resources. It is likely that theinclusion of real resources along with dollarswill increase the cost of the student-levelresource measurement system, and whilereal resources probably have a slightly lowerpriority than dollar measures, we recom-mend that they be considered for inclusion.

•Variable versus fixed costs. Once therelationship between costs and outcomes aredetermined, from policy and managementperspectives, it is useful to know whether thecosts are variable or fixed. However,because the distinction between fixed andvariable costs is dependent on the time frameused to assess costs, we deem this distinctionto be a second order priority. Also, althoughdirect (indirect) costs are not always variable(fixed), they often are, and analysts may beable to approximate variable (fixed) costs byusing selected items of direct (indirect)costs.

Conclusions

NCES should move ahead with thedevelopment of a framework for a student-level resource measure. Whether or notNCES actually collects the data needed todetermine student-level resource measureswill depend on an agency-level cost benefitanalysis. We believe that the case is strongenough for the necessary development toproceed and make the following recommen-

dations:

1.Student-level resource measuresshould be designed primarily to answerquestions of resource effectiveness andsecondarily to answer questions ofequity and intent. When conflicts arise,NCES should choose the resourceeffectiveness goal because questions ofproductivity are the most pressing forthe public, policy makers, and research-ers.

2.Development of a system to measurestudent-level resources should build onother NCES efforts such as financialaccounting frameworks and longitudinaldata bases.

3.Resource measures should be linkedto student outcomes.

4.Measures should be student-based(product as opposed to departmental)costing ones and should include full notpartial costs.

5.The measures should include directand indirect costs, with detailed break-downs in each category.

6.Measures that include costs of educa-tional programs are a high priority.

7.There should be heavy reliance on joborder costing where appropriate.

8.If possible, there should be indicationsof alternative funding sources (federal,state, and local).

9.NCES should include the methodsused to allocate indirect costs in thedescription of the resource measures.

10.The measures should distinguishbetween operating and capital costs.

Whether or notNCES actually

collects the data

needed todetermine

student-level

resourcemeasures willdepend on an

agency-level costbenefit analysis.


11.If possible, NCES should includemeasures of real resources as well as thedollar value of resources.

12.The measurement of fixed versusvariable costs is a lower priority.

A student-level resource measure can beadopted to one of NCES’s existing efforts,for example NELS:88. One of the crucialissues is whether the costing methodologyneeds to be carried out for every student inthe district, because of the nature of costallocations, or whether accurate costingmethods can be developed based on asample similar to the one in NELS:88. Webelieve that a sampling method is possible.

Finally, we want to emphasize theimportance of developing a good set ofstudent-level resource measures to accom-pany NCES data bases such as NELS:88. Itis crucial that we begin to make progress onthe question of effective use of resources ineducation. The NELS:88 data base iscertainly one of the best existing data basesfor researching that question, but without anappropriate set of resource numbers thesedata will go unused for this purpose.

It is crucial thatwe begin to makeprogress on the

question ofeffective use ofresources in

education.


References

Production Function Studies

Brown, B.W., and D.H. Saks. 1987. “The Microeconomics of the Allocation of Teachers’ Time and Student Learn-ing.” Economics of Education Review 6: 319-332.

Ferguson, R.F. Summer 1991. “Paying for Public Education: New Evidence on How and Why Money Matters.”Harvard Journal on Legislation 28: 465-498.

Hanushek, E.A. 1986. “The Economics of Schooling: Production and Efficiency in Public Schools.” Journal ofEconomic Literature 24: 1141-1177.

King, R.A., and B. MacPhail-Wilcox. Summer 1994. “Unraveling the Production Equation: The Continuing Questfor Resources that Make a Difference.” Journal of Education Finance 20: 47-65.

Monk, D.H. Winter 1992. “Education Productivity Research: An Update and Assessment of Its Role in EducationFinance Reform.” Educational Evaluation and Policy Analysis 14: 307-332.

Equity Studies

Berne, R., and L. Stiefel. Winter 1994. “Measuring Equity at the School Level.” Educational Evaluation and PolicyAnalysis 16: 405-421.

Picus, L.O. March 1993. “The Allocation and Use of Education Resources: School Level Evidence from theSchools and Staffing Survey.” Center for Research in Education Finance, Working Paper 37.

Resource Intent Studies

Chambers, J., T. Parrish, M. Goertz, C. Marder, and C. Padilla. January 26, 1993. “Chapter 1 Resources: Supple-menting an Equal Base?” Final Report: final draft, submitted to U.S. Department of Education.

Cooper, B., and Associates. “Making Money Matter in Education: A Micro-financial Model for Determining School-Level Allocations, Efficiency, and Productivity.” Journal of Education Finance 20: 66-87.

Coopers & Lybrand L.L.P. October 4, 1994. “Resource Allocations in the New York City Public Schools, draft.” onbehalf of Mayor Rudolph W. Guiliani and Herman Badillo, Esq.

Levin, H. 1983. Cost-Effectiveness: A Primer. Volume 4. Sage Publications, Beverly Hills, CA.

Parrish, T.B. Winter 1994. “A Cost Analysis of Alternative Instructional Models for Limited English ProficientStudents in California.” Journal of Education Finance 19: 256-278.

Other References

Finkler, Steven A. 1994. Cost Accounting for Health Care Organizations: Concepts and Applications.Gaithersburg, MD: Aspen Publishers, Inc.


Fowler, William J., Jr. 1990. Financial Accounting for Local and State School Systems, 1990. U.S. Department ofEducation. Office of Educational Research and Improvement. Washington, DC: National Center for EducationStatistics.

U.S. Department of Education. Office of Educational Research and Improvement. National Center for EducationStatistics. March, 1990. National Education Longitudinal Study of 1988. User’s Manual. Washington, DC: NCES90-482.

Proposed “Good Practices”for Creating Data Bases fromthe F-33 and CCD for School

Finance Analyses

Selected

Papers in

School

Finance


Proposed ”Good Practices” for CreatingData Bases from the F-33 and CCD

for School Finance Analyses

Michael O’Leary and Jay MoskowitzWashington, DC

About the AuthorsMr. O’Leary is a Senior Associate in

Coopers & Lybrand’s Education and Non-Profit Consulting Practice, currently special-izing in administrative restructuring ininstitutions of higher education. Prior to hiswork at Coopers & Lybrand, he was aResearch Analyst at Pelavin Associates inWashington, DC. While at Pelavin Associ-ates, his work focused on K–12 schoolfinance and the Fiscal Crosswalk Project.He also managed a project to assess equityand revenue distribution among the nation’s15,000 public school districts.

He is co-author of A Comparison ofState and Federal School Accounting

Procedures (National Center for EducationStatistics, U.S. Department of Education,and Governments Division, Bureau of theCensus, 1992), Trends in EducationalExpenditures (American Educational Fi-nance Association, Annual Conference,1992), and An Assessment of Training andFacilities Needs within Tribally ControlledPostsecondary Vocational Institutions(Report to Congress, U.S. Department ofEducation, 1993).

Mr. O’Leary is a graduate of DukeUniversity where he also served as a memberof the Board of Trustees from 1991–94.

Jay Moskowitz, Managing Associate,Pelavin Research Institute of the AmericanInstitutes for Research, conducts andmanages international and comparativeeducation projects. Current assignmentsinclude serving as staff director for OECDindicators network measuring studentoutcomes, preparing comparative casestudies of teacher induction practices amongAPEC members, developing methods forinternational benchmarking of educationstandards, and managing the development ofan APEC member dialogue on performancemeasurement systems. Prior to his recentfocus on international education, Dr.Moskowitz worked extensively with states

and federal agencies on issues of schoolfinance and equity.

He is co-author of Plain Talk aboutSchool Finance, Targeting Formula, andResource Allocation Issues: FocusingFederal Funds Where the Needs are Great-est, Fiscal Equity in the United States 1984–85, Research in School Finance: A concernfor Equality and Equity, Interpreting Changein Educational Equity, and A School BasedAnalysis of Inter- and Intradistrict ResourceAllocation.

Dr. Moskowitz received his Ph.D fromthe Maxwell School, Syracuse University.

Proposed “Good Practices” for Creating Data Bases 85

At least three studies1 sponsored by theU.S. Department of Education utilized thesame school finance and enrollment data setsas their sources—the Survey of LocalGovernment Finances for School Systems(F-33) and Common Core of Data (CCD).Realizing the need for common data basesupon which to conduct their analyses,researchers working on these studies at theAmerican Institutes for Research (AIR),Consortium for Policy Research inEducation’s Finance Center (CPRE), andPelavin Associates² cooperated to develop aset of common practices regarding data basecreation. While the researchers’ ultimategoals and methodologies were different, theywere able to create nearly identical databases by collaborating and adopting eachother’s techniques for selecting appropriate

Introduction

Background

With recent developments in schoolreform and state and Federal financing ofpublic schools, the desire for accurateschool finance (and related enrollment andprogrammatic) data has grown. The wide-spread use of personal computers which canrun sophisticated statistical software pack-ages and analyze very large data sets hasalso made it much easier for researchers toscrutinize public data sets. Computer runsthat used to take hours can now be done inminutes or seconds, allowing researchers tosubject data bases to new levels of scrutiny.The increased scrutiny certainly improvesthe quality of the data, but it also opens thedoor to wide variations in the proceduresused by individual researchers (even thosein the same fields) to create data bases foranalysis (from multiple, larger data sets).

Proposed “Good Practices” for CreatingData Bases from the F-33 and CCD forSchool Finance Analyses

Michael O’Leary and Jay MoskowitzWashington, DC

1 Parrish, T.B., et al. 1995; Hertert, L. 1994; and O’Leary, M., et al. 1995.

2 After the initiation of these studies, Pelavin Associates, Inc. became an affiliate of the American Institutes for Researchand was renamed Pelavin Research Institute.

With recentdevelopments inschool reform

and state andFederal financingof public

schools, thedesire foraccurate school

finance...data hasgrown.


proposed process includes the steps taken toaddress the issues identified by AIR, CPRE,and Pelavin Associates and notes theadditional issues (e.g., adjustment forspecial-needs pupils) that could not beaddressed fully or were not common to allthree studies.

Major National Data SetsUsed For School FinanceResearch

The Census Bureau’s Survey of LocalGovernment Finances for School Systems(F-33) and the National Center for Educa-tion Statistics’ (NCES’) Common Core ofData (CCD) are the two national data setsmost relevant to school finance research.

The F-33 is conducted annually by theCensus Bureau under contract with NCES.Although conducted annually, it is not basedon a universe of school districts each year.Typically, the survey includes all districts inall states in years ending in 2 or 7 (1982,1987, and 1992). To obtain financial datawhich corresponded more closely to thedecennial censuses, NCES also sponsoredspecial F-33 surveys in 1980 and 1990.

Consequently, F-33 data are availablefor all districts for the 1979–80, 1981–82,1986–87, 1989–90, and 1991–92 schoolyears. For the intervening years, the F-33includes data on all districts for 33 to 35states, but only a sample of districts in theremaining 15 to 17 states.

The CCD contains data at three differ-ent levels from five different surveys. It hasthree separate files for state, agency (dis-trict), and school-level data. These files arebased on data from the F-33 and the Schoolsand Staffing, Public Elementary-SecondaryAgency, Public Elementary-SecondarySchool, State Non-Fiscal Elementary andSecondary Education, and National Public

district records from their primary datasources: the F-33 and CCD surveys (rev-enue, expenditure, and enrollment data) andthe District-Mapped Decennial Census(demographic and economic data). Revenue,expenditure, and enrollment data from the F-33 and CCD surveys are the sine qua non fornational school studies and demographic andeconomic data.

Objectives

This paper provides future users of thesedata sets (and subsequent releases thereof)with an introduction to the major issuesregarding creation of a data base for analy-sis. Based on the procedures followed bythe AIR, CPRE, and Pelavin Associatesresearchers, this paper summarizes a meth-odology for selecting the records (i.e., schooldistricts) appropriate for most school financeanalyses.³ This methodology is offered as apotential model for future data base creationand modification.

By presenting the issues and caveatsregarding initial modification of the data,and by proposing a methodology for database creation, the authors hope the paperwill encourage greater standardizationamong data users. Individual researchmethods will (and should) always vary, butwith common data base creation procedures,researchers’ conclusions can be debated ontheir merits rather than dismissed due tovariations in the underlying data sets.

Organization

Subsequent sections of this paperinclude descriptions of the F-33 and CCDdata sets; the use of these data sets in paststudies; issues identified during creation ofthe data bases for the AIR, CPRE, andPelavin Associates studies; and a proposedprocess for modifying the data sets. The

3 Readers interested in the specific procedures used in the AIR, CPRE, and Pelavin Associates studies are encouraged toconsult with the authors of those studies directly.

Individualresearch methodswill (and should)

always vary, butwith common

data base

creationprocedures,researchers’

conclusions canbe debated on

their merits

rather thandismissed due tovariations in the

underlying datasets.


Education Financial Surveys.

The District-Mapped Census data areavailable as a result of the NCES-sponsored“School District Mapping Project.” For theproject, the Census Bureau used household-level data by census tract from its 1980 and1990 decennial censuses. The Bureau tookgeographic coordinates for each state’scensus tracts and school districts andaggregated (or mapped) the tract-level dataup to the district level. As a result, theDistrict-Mapped Census data include a widevariety of demographic and socioeconomicinformation for each of the nation’s schooldistricts.

Recent Uses Of Major SchoolFinance Data Sets

The F-33 and CCD data sets have beenused extensively for studies conducted byresearchers at AIR, CPRE, and PelavinAssociates during the last two years. TheAIR and Pelavin Associates researchers alsoused some of the district-mapped data. Thestudies’ foci, or emphases, are presented inTable 1 and the data elements used by thePelavin Associates study are presented inTables 1a and 1b.

Other major national school financestudies utilizing the F-33 data include thoseby Wyckoff (1992), Riddle (1990), andSchwartz and Moskowitz (1988). Wyckoffused enrollment and current and instruction-al expenditure data from the F-33 to exam-ine changes in intrastate equity between1980 and 1987. Riddle also examinedintrastate equity using F-33 data, but for1987 only. Schwartz and Moskowitz usedF-33 enrollment, expenditure, and revenuedata, as well as 1980 district-mappedincome, poverty, and property data toexamine changes in horizontal equity andequal opportunity between 1977 and 1985.

Unlike the AIR, CPRE, and PelavinAssociates studies, these three earlier studiesdid not use a common approach to dataprocessing and provided only summaryinformation about the approach that they didtake. For example, Schwartz and Moskowitzdiscussed the problems associated with usingsample data and listed the number of dis-tricts included in their analysis (by state), butthey did not indicate what criteria they usedto include those districts and exclude others.

Riddle included a lengthy discussion ofthe limitations of school finance data, butprovided very little information about which(or how many) districts he included in hisdatabase. Other than noting that he screenedout districts with low enrollments,4 heprovided no information about the creationof his database and no information aboutwhether he screened for special or non-operating districts.

Wyckoff provided more detail about hisdata base than Riddle, Schwartz, orMoskowitz. Like Riddle, Wyckoff notedsome of the basic limitations of the data(e.g., state by state variations in classifyingexpenditures), but Wyckoff also noted thetotal number of districts in the 1980 and1987 data sets, the number eliminatedbecause they were not unified, and thenumber eliminated because they werecommunity college, vocational, non-operat-ing, or service districts. However, Wyckoffdid not provide the number of districtsincluded by state or any other details abouthis data base and the variables containedtherein.

Given this lack of information andapparent inconsistency regarding dataprocessing, researchers working on the AIR,CPRE, and Pelavin Associates studiescollaborated and, to the extent possible,standardized their data base creation process.

4 “To avoid marginal cases where expenditures per pupil are very high largely because the [districts’] enrollment is quitesmall,” Riddle eliminated unified districts with fewer than 500 students and non-unified districts with fewer than 250students.

The F-33 andCCD data setshave been used

extensively forstudiesconducted by

researchers...


Table 1.—Foci of three recent school finance studies

Study Focus

AIR Variations in school district revenues andexpenditures between districts of differingcharacteristics (e.g., size, urbanicity,geographic region, and student composition)in 1990

CPRE (1) Magnitude and patterns of equity inpublic school revenues by state and region in1990; and(2) Cost of equalizing revenues within statesin comparison to the cost of equalizingrevenues by region or across all states

Pelavin (1) Trends in revenue-based school finance Associates indicators within states between 1980 and

1992; and;(2) State profiles of school finance indicatorslegislation, and litigation

SOURCE: Pelavin Associates.

Table 1a.—Census variables used by Pelavin Associates

1980 1990 Survey Cell Description Table Item No. Table Record Type Item No.

Median Household Income 69 1 P080A 1 001

Residential Property Value 140* 1 H062 1 001, 002

Percentage of Children 94 8 P017B 7 003 to 008 in Poverty

*Scaled by 250,000.



Table 1b.—F-33 Variables used by Pelavin Associates

1980 1987 1990 1992 Variable Variable Variable Variable

Survey Cell Description Code Code* Code* Code*

Local Property Taxes T01 T06 T06 T06 Local Gen. Sales Gross T09 T09 T09 T09 Receipts Taxes Local Public Utility Taxes T15 T15 T15 T15 Local Personal Income T49 T40 T40 T40 Local Corporate Net Income T41 T41 Taxes Local Parent Govt. Contribs. T02 T02 T02 T02 Local Revenue - Cities & D21 D23 D23 D23 Counties Local Gross Receipts From A09 A09 A09A09 School Lunch Local Other Sales & Service A12 A12 A12A11, A13, Revenue & A20 Local Misc. Other Local U99 U99 U99U97 & Revenues C24 Local All Other Taxes T99 T99 T99 T99 Local Education, Other Charges A21 A21A21 &

A15 Local Interschool Transfer D11 D11 D11 D11 State Direct Aid C21 C23 C23 C01, C04-

C13, C35 State Revenues on Behalf C27 C27 C38 & C39 of Schools Federal thru ESEA C22 C26 C26 — Federal thru NDEA C23 C26 C26 — Federal Child Nutrition C25 C25 C25 C25 Federal All Other C26 C26 C26 C14-C20,

C36, B11,B12

Federal Direct P.L. 815 B23 B21 B21 B10 Federal Direct P.L. 874 B24 — — — Federal Direct All Other B26 — — B13 Enrollment E57 V33 V33V33 Current Expenditures E11-E17 E12 E12 E12, V70,

V75, V80

* Every 1987, 1990, and 1992 variable, except Enrollment (V33), was scaled up by a factor of 1,000

to obtain comparable values.

— Not available.



The issues identified during the course ofthis collaboration are discussed in thefollowing section.

Data Base Creation Issues AndProcess

While the F-33 and CCD data sets aretremendously valuable resources, they havetheir limitations, and there are a number ofdata base creation issues which must beaddressed clearly by any users of the data.These issues deal primarily with observationor record selection—i.e., the process bywhich “appropriate” districts are selected foranalysis and inappropriate districts arewinnowed out.

Three key issues emerged as the re-searchers planned their studies and createdtheir data bases:

1. F-33 Sampling,

2. District Types, and

3. District Levels.

Each of these issues and the standard-ized process used to address them is dis-cussed below.

F-33 Sampling

For most years, the F-33 survey is basedon a sample of districts in 15 to 17 states.Realizing this, and adjusting or limitinganalyses accordingly, are crucial first stepsfor any study using the data. One researcherdevoted considerable effort to examining“the absence of revenue data for manydistricts,” without realizing that he was usingF-33 data from a sample year (Toenjes1994).

In the past, researchers have dealt withthis limitation in a variety of ways. Forexample, Schwartz and Moskowitzweighted data for 17 states5 for their 1984–85 equity analysis. They note, “It is diffi-cult to assess the impact of these samplingprocedures on... estimates of fiscal equity[because] probability distributions of equitystatistics are either unknown or are difficultto obtain, and....the sampling procedures arequite complex and variance estimates arenot straight-forward.”

In part to parallel the district-mappeddata, the AIR study used data for 1990only—also an F-33 universe year. Giventhe F-33 sampling issues, the PelavinAssociates and CPRE studies also used F-33data for universe years only. Researchersinterested in the intervening years eitherhave to (1) limit their study to the 35 statesfor which data from a universe of districtsare available, or (2) tolerate the uncertainlevel of error resulting from weighted datafor 15 to 17 states, or (3) attempt to estimatethe sampling error for the weighted data.

District Types

The F-33 and CCD contain varioustypes of districts which are “special”because of their instructional programs orstudents. These special districts includecommunity college, agricultural, vocational-technical, correctional/custodial, specialeducation (e.g., BOCES districts in NewYork State), and “non-operating” districts.Non-operating districts have students but nobuildings and include regional supportcenters, co-ops, media centers, and othersupervisory or service districts.

The AIR, CPRE, and Pelavin Associatesstudies went to considerable lengths toexclude such districts because: (1) theirprograms have considerably different

5 Alabama, Arizona, Arkansas, Colorado, Georgia, Indiana, Kentucky, Mississippi, Montana, New Jersey, New Mexico,Oklahoma, Oregon, South Carolina, South Dakota, Utah, and Vermont.

...the F-33 andCCD data sets

are tremendously

valuableresources...


resource needs; (2) they often receive statefunds through categorical aid formulas thatare different from those used for “regular,operating” districts; and (3) their bordersoften overlap or encompass those of regular,operating districts.

For studies that include comparisons ofcurrent expenditures, non-operating districtscan present additional problems. States likeMassachusetts, Nebraska, and Vermontreport administrative and support expendi-tures in special administrative, supervisoryunion, or service districts. As a result,function-level expenditure data for districtsin these states would underestimate the truecosts of school administration unless thenon-operating districts expenditures wereallocated back to the operating districts thatthey served. Currently, the codes in the F-33 data base do not indicate the relationshipbetween operating and non-operatingdistricts. To accurately include all expendi-tures by function, operating and non-operating districts would have to bematched by hand and then a formula wouldhave to be created to allocate the adminis-trative and support service expendituresreported by the non-operating districts backto the operating districts they served.

However, the F-33 does include specialcodes for regular, operating, and “special”districts. These codes are shown in Table 2.The AIR, CPRE, and Pelavin Associatesstudies all used these codes as an initialscreen of their data and eliminated morethan 1,200 observations.

Second, they screened for districts withzero or missing enrollment and expendituresor revenues. Districts with no enrollment

may have been non-operating or may simplyhave been errors in the original creation ofthe original data base. In either case,enrollment and the F-33 revenue/expendituredata were the sine qua non for these studies.

Third, the researchers used text-searchesto flag observations with “VOC,” “TECH,”“SPEC,” or “AGRIC” in the district namefield. These districts were then reviewed byhand to ensure that, for example, “spec” wasnot found in the middle of a regular district’sname.

Fourth, they used individualized educa-tion plan (I.E.P.) counts aggregated from theCCD school file to the district level andscreened out any districts with more than 50percent I.E.P. or “special education” stu-dents.

Fifth, they used the CCD district typecodes (Table 3) for a final screen. Thesedistrict type codes identify supervisoryunions, regional service districts, state-operated districts (e.g., districts for handi-capped students or students who are in thestate’s juvenile correctional system), andfederally-operated districts (e.g., Departmentof Defense schools). Combined, these fivestages of screening eliminated or “winnowedout” nearly 3,000 observations that were notregular, operating districts.6

District Levels

There are four basic “levels” of districts:(1) elementary-only, (2) secondary-only, (3)unified K–12, and (4) college-graded.7 Formost school finance studies, the college-

...the F-33 doesinclude specialcodes for regular,

operating, and“special”districts.

6 Note that these CCD codes are actually in the TYPECODE field, while the F-33 codes are in the SCHLVLCOD field.Note also that the SCHLVLCOD field encompasses both district levels (i.e., elementary-only, secondary-only, and unified)and district types (i.e., operating, non-operating, vocational, etc.).

7 Note that the terms “level” and “type” are used inconsistently in the literature. For this paper and the Pelavin Associatesstudy, “level” refers to the grade levels included in a district and “type” refers to the variations of regular andspecial, operating and non-operating districts.


Table 2.—F-33 district code numbers used to screen for regular, operating districts

Year District Type 1980, 1982 1987 1990, 1992

Operating 1 1 (elementary), 1,2,3 (same 2 (secondary), as 1987) 3 (unified)

Non-Operating 2 6 6

Supervisory Union 3 — —

Operating system which — — is a component of a supervisory union

Non-operating system 5 — — which is a component of a supervisory union

Public agency for special 6 5 & 7 5 & 7 education, voc/tech

College-grade system — 4 —

— Not available.


Table 3.—CCD codes used to screen for regular, operating districts

District Type Code

Regular 1 and 2 Operating Supervisory union administrative centers 3 Regional education service agencies 4 State-operated agencies 5 Federally operated agencies 6 Other agencies 7



graded districts can be winnowed outimmediately. They are typically communitycolleges which are not directly relevant tostates’ support of elementary and secondaryeducation.

Over the years as states have encouragedmergers and consolidations of districts, thedistribution of districts among the first threelevels has shifted toward unified districts.However, there are still some states with asignificant number of non-unified districts,and there are also two states (Montana andVermont) with no unified districts. Unlikethe earlier study in which Wyckoff elimi-nated all non-unified districts (approxi-mately 25 percent of his data base for 1987),the AIR, CPRE, and Pelavin Associatesstudies included all three levels of districts.However, they handled the analysis thereofsomewhat differently.

As Wyckoff, Riddle, and others havenoted, past research has demonstrated thatcosts vary by district levels. Based partiallyon arguments made by Odden and Picus inSchool Finance: A Policy Perspective, theCPRE study standardized non-unifieddistricts’ revenues to those of unifieddistricts by increasing elementary-onlydistricts’ revenues by 10 percent and de-creasing secondary-only districts’ revenuesby 25 percent.

The AIR and Pelavin Associates studiescombined all district levels and relied onpupil-weightings to mitigate the effects ofnon-unified districts (which tend to havevery low enrollments). Representing a thirdapproach to the district level issue, Riddleanalyzed financial indicators separately foreach level.

Whether weighting, separate presenta-tion, or elimination is chosen as the appro-priate way to adjust for district levels,researchers can utilize the SCHLVLCODfield in the F-33 (see Table 2) and theGRADESPAN field in the CCD for yearssince 1985–86. For districts with missingvalues in the SCHLVLCOD field, AIR,CPRE, and Pelavin Associates used valuesfrom the GRADESPAN field. Then, theresearchers used text searches of the dis-tricts’ names and manual checks for districtswhich still had missing level codes.8

Unstandardized or Unresolved DataBase Creation Issues

While the AIR, CPRE, and PelavinAssociates researchers developed a standardprocess for choosing the districts forinclusion in their data bases, they tookslightly different approaches to or were notable to fully address issues in the followingareas:

• Enrollment

• Special-needs pupils

• Property, poverty, and income data

• State and district finances.

Researchers using the F-33, CCD, anddistrict-mapped data should be aware ofthese issues as they develop their own databases and study designs.

Enrollment

The F-33 and CCD both include enroll-ment variables. The AIR and Pelavin

8 To obtain district types for all districts prior to 1986–87, Pelavin Associates researchers merged the 1980 district IDcodes with the 1987 codes. For districts which did not match, the researchers used text searches of the district namesand manual checks. As a result, the 1980 district level codes are “best guesses.” It is possible that changes in districtlevels occurred between 1980 and 1987 which would have been obscured by the Pelavin Associates ID code merge. Thisis just one example of problems associated with use of the 1980 data. Additional difficulties encountered by the PelavinAssociates researchers included missing or incomplete poverty and property data for several states (e.g., Alabama andCalifornia), and incomplete documentation of the source data sets’ record layouts.

Over the years asstates haveencouraged

mergers andconsolidations ofdistricts, the

distribution ofdistricts...hasshifted toward

unified districts.


egorical funds (from any source) cannot beexcluded. The Pelavin Associates studyexamines the statistical impact on basicschool finance indicators using revenues orexpenditures.

Until complete and nationally compa-rable data identifying the number and typesof special-needs students in each district andthe amount of funds allocated through basicand categorical funds are available, schoolfinance researchers will struggle to disen-tangle the effects of variations in studentneeds and district circumstances and thesubsequent variations in the allocation offunds to districts. In the meantime, theCCD data and District-Mapped Censusprovide some proxies.

The CCD school file includes data onurbanicity, race/ethnicity, special education(I.E.P.) status, and free and reduced pricelunches. The District-Mapped Census alsoincludes a variety of demographic andsocioeconomic data (e.g., race/ethnicity,limited English proficiency, and poverty).Adjustments based on these data, in con-junction with the unadjusted data, will shedsome light on the relationship between pupilneeds and school finances, but much moreresearch needs to be done regarding theactual range in (and adequacy of) fundingfor special needs students and districts.

Property, Poverty, and Income

In order to examine the relationshipbetween the wealth or affluence of a schooldistrict and its revenues and expenditures,the Pelavin Associates study utilizedDistrict-Mapped Census data on home-owner estimates of residential propertyvalue (in lieu of the more commonly usedassessed or equalized property valuation perpupil which is not available on a nationally-comparable basis), child poverty counts, andmedian household incomes. Real property(upon which taxes are levied) includes

Associates studies used the F-33 enrollmentsand relied on the CCD to fill in missingvalues where possible. The CPRE studyused the CCD enrollments.

For 1990, CCD enrollments exceeded F-33 enrollments by 2.5 percent or more for 10percent of districts. At the other end of thescale, F-33 enrollments were at least 2.5percent greater than CCD enrollments for 25percent of all districts. Based on manualchecks of districts with large discrepanciesbetween CCD and F-33 enrollments, neithersurvey appears to be “correct.” Commoncauses of the discrepancies appear to be (1)separate reporting of elementary and second-ary schools on one survey and consolidatedreporting on the other, and (2) miscoding ofdistricts with the same names in the samestate. The effects of these variations on per-pupil school finance indicators are uncertainand warrant additional study.

Special-Needs Pupils

Given the existing F-33 and CCD data, itis impossible to isolate all special districts,schools, and students and their respectivefinancial data. Some special-needs (e.g.,special education) districts were eliminatedbased on the I.E.P. counts and district typecodes from the CCD. However, it is impos-sible to separate funds targeted for special-needs pupils and special-circumstancesdistricts (e.g., geographically isolated ones)from those allocated through basic aidprograms.

The AIR and Pelavin Associates studiesused state and local revenues in order toeliminate Federal categorical funds fromtheir calculations. This provides a moreaccurate picture of the fiscal disparitiesinherent to a state’s funding system but aless comprehensive picture of fundingavailable to serve special-need pupils wherefederal funds play an important role. Foranalyses using district expenditures, cat-

Untilcomplete...dataidentifying the

number...ofspecial-needs

students in each

district and theamount of funds

allocated...are

available, schoolfinance

researchers will

struggle todisentangle the

effects of

variations instudent needs...

district

circumstancesand...subsequentvariations in the

allocation offunds to districts.


commercial, industrial, and agriculturalproperty as well as residential property.Because residential property includes only asmall portion of the real tax base in manydistricts and is distinctly different fromassessed property valuations, the applicabil-ity of these data to school finance studieswithout other measures of affluence isquestionable. Property data were alsounavailable for a large number of districts,especially in California. Data for thesedistricts had to be estimated using growthrates for property values in surroundingdistricts.

More comprehensive and appropriateproperty data are not currently available forall states. Any study using the District-Mapped Census data on residential propertyvalues must acknowledge all of theseshortcomings.

Given the limitations of the propertydata, child poverty counts and medianhousehold income provide more accurateestimates of districts’ fiscal capacity.However, they too have their limitations.For example, a district’s fiscal capacitycould be high in terms of real propertyvalues even though it had high child povertyrates or low median incomes. The District-Mapped Census data represent a major steptowards better analysis of the relationshipbetween fiscal capacity and school finance,but conclusions based on them must still bemade with great caution.

State and District Finances

Researchers using the F-33 data mustdeal with a variety of issues regarding stateand district finances. Because they haddifferent goals and used different financialvariables, the AIR, CPRE, and PelavinAssociates studies did not use the sameapproach to the following issues:

• On-Behalf-of-LEA Funds—In moststates, the total on-behalf-of-LEA revenues

and expenditures of districts in the F-33 datafile are not equal to the amount of DirectState Support provided by the states. TheAIR study developed an imputation proce-dure to allocate all funds to the appropriatedistrict accounts.

• “Outliers”—All three studiesscrutinized districts with per-pupil expendi-tures and revenues in the top and bottom 1 or0.5 percent of the distribution. Determina-tions about whether to include or excludethese outlier districts were made based onmanual checks and contacts with Census,state, and district officials. These determina-tions were not necessarily standard acrossthe three studies.

• Fund Imbalances—Considerabledifferences between total revenues andcurrent expenditures exist. In some cases,this is due to districts’ financing capitalexpenditures from their general revenuefunds; in others, it is due to intergovernmen-tal transfers (see below). The studies werenot able to fully address these cases.

• Intergovernmental Transfers—Thenumber of pupils covered by intergovern-mental transfers is not apparent from the F-33 or CCD data. As a result, per-pupilexpenditures and per-pupil revenues may bequite different. Adding intergovernmentaltransfers to operating expenditures becausethe transfers may contain more than theinstructional costs for children attendingschools in other districts. While this issuecould not be “resolved,” the studies followedNCES practice and did not include transfersin their expenditures.

Summary

Despite the fact that the AIR, CPRE, andPelavin Associates researchers were not ableto resolve all the issues regarding data basecreation, their collaboration and the commonprocess they used greatly increased thecomparability among the data bases they

Researchers usingthe F-33 datamust deal with a

variety of issuesregarding stateand district

finances.


ultimately used for their analyses. Conse-quently, these studies can be compared andevaluated on the basis of their results andanalytical methods alone to a much greaterextent than previous studies which variedwidely in their data base creation proce-

Table 4.—Proposed “best practice” process for selecting appropriate districts for ananalysisdata base

File Merge Merge CCD and F-33 files to replace missing enrollments | | Step 1 Winnow out special or non-operating districts based on the F-33 district types | | Step 2 Winnow out districts which still have zero or missing enrollments or have zero or missing revenues and expenditures | | Step 3 Winnow out districts with VOC, TECH, SPEC, or AGRIC in their names and then hand check the districts to be excluded | | Step 4 Winnow out districts with greater than 50 percent of their enrollment classified as special education based on CCD I.E.P. counts | | Step 5 Winnow out districts based on CCD district type codes | | Step 6 Winnow out districts based on F-33 and CCD district level and grade-span codes


...these studiescan be comparedand evaluated on

the basis of theirresults andanalytical

methods alone toa much greater

extent...

dures.

The common process used is summa-rized in Table 4 as a model or suggested“best practice” for other researchers usingthe F-33 and CCD data. Once the CCD andF-33 are merged and missing enrollmentsare filled, the order of the steps in thiswinnowing process is not crucial.


References

Hertert, L., Busch, C., and Odden, Allan. Winter, 1994. “School Financing Inequitites Among the States: TheProblem from a National Perspective.” Journal of Education Finance 19(3): 231-255.

O’Leary, M., Moskowitz, J., and O’Malley, Amy. 1995. Trends in School Finance Equity Indicators, 1980-1992.Washington, DC: Pelavin Research Institute. unpublished manuscript.

Parrish, T.B., Matsumoto, C.S., and Fowler, W.J., Jr. 1995. Disparities in Public School District Spending 1989-90:A Multi-variate, Student-weighted Analysis, Adjusted for Differences in Geographic Cost of Living and Student Need.Washington, DC: National Center for Education Statistics, U.S. Department of Education.

Riddle, Wayne Clifton. 1990. “Expenditures in Public School Districts: Why Do They Differ?” Washington, DC:Congressional Research Service.

Schwartz, Myron, and Jay Moskowitz. 1988. “Fiscal Equity in the United States, 1984-85.” Washington, DC:Decision Resources Corporation.

Toenjes, Laurence O. March, 1994. “Interstate Revenue Disparities and Equalization Costs.” Presented at theAnnual Conference of the American Education Finance Association in Nashville, Tennessee.

Wyckoff, James H. 1992. “Intrastate Equality of Public Primary and Secondary Education Resources in the U.S.,1980-87.” Economics of Education Review 11(1): 19-30.

The Empirical Argument forEducational Adequacy, the CriticalGaps in the Knowledge Base, and

a Suggested Research Agenda

Selected

Papers in

School

Finance


William H. Clune is Voss-BascomProfessor of Law at the University ofWisconsin Law School, Director of thePolicy Group of the National Institute forScience Education, and a senior researcherof the Consortium for Policy Research inEducation (CPRE). His past research hasincluded school finance, school law, imple-mentation, special education, public em-ployee interest arbitration, school site

The Empirical Argument forEducational Adequacy, the CriticalGaps in the Knowledge Base, and a

Suggested Research Agenda

William H. CluneUniversity of Wisconsin-Madison

autonomy, effects of high school graduationrequirements, upgrades of the high schoolcurriculum in mathematics and science, andsystemic educational policy. His presentresearch includes “program adequacy” (thecost and implementation structure needed toreach high minimum levels of studentachievement in low-income schools) andsystemic policy in mathematics and scienceeducation.

About the Author

The Empirical Argument for Educational Adequacy 103

1994b; Levin 1991); and remedial educationis the target of especially strong skepticismand political resistance (Goertz 1983;Herrnstein and Murray 1994).

Under these circumstances, in which anemphasis on outcomes and skepticism aboutcosts exists, it is natural that strains aredeveloping in the traditional goal of equityin school finance (Clune 1994b). Equityguarantees inputs regardless of outcomes ata time when social pressures are moving inthe opposite direction, toward outcomeswhich are guaranteed and costs which arecontingent. This paper responds positivelyto the social environment by trying todevelop a research agenda which has thecapacity to clarify the relationship betweeneducational investments, educational prac-tice, and educational outcomes for thedisadvantaged. The fundamental goal iscost-effectiveness, or productivity (Clune1995c). As a later section will argue, oncesociety (including the courts) becomessatisfied with a vision of how high minimumeducational outcomes can be produced atminimum cost within an affordable budget,

Introduction

This paper develops the idea that thegoal of educational adequacy, that is, highminimum educational outcomes for thedisadvantaged is constrained by lack of astrong knowledge base and that a well-formulated research agenda can make animportant contribution to policy.

Currently, school finance is subject toconflicting and confusing pressures. On onehand, improved educational outcomes haveprobably never been so highly valued. Thelink between successful schooling andeconomic prosperity for both the individualand society is widely accepted (Gamoran1994; National Education Association1995). On the other hand, the link betweeneducational outcomes and educationalreform and investments is widely doubted(Hanushek 1991), and resources for educa-tion and all other public spending arebecoming more limited (Gold 1995).Education for the disadvantaged raises theseconcerns in a particularly acute way. Poorchildren make up the bulk of students whoseeducational outcomes are seriously beloweconomically functional minimums (Berne

The Empirical Argument for EducationalAdequacy, the Critical Gaps in theKnowledge Base, and a SuggestedResearch Agenda

William H. CluneUniversity of Wisconsin-Madison

...high minimumeducational

outcomes for thedisadvantaged isconstrained by

lack of a strongknowledge base...


misery), but questions of equality are evenmore obviously controversial and contingentupon available wealth.

Problems of defining adequacy mightseem to justify research on the social costsand benefits of education for the disadvan-taged. Henry Levin, for example, developeda case for compensatory spending based onboth the benefits of good education and thesocial costs of poor education (1991). Thisand other arguments about returns toeducation are certainly relevant, but thesocial controversy surrounding the issuemakes them necessarily inconclusive (e.g.,how certain are we that compensatoryspending in the present will reduce crimecosts in the future, and how willing aretoday’s taxpayers to make such an invest-ment?).

Fortunately, defining adequacy is one ofthose rare problems which is practicallymanageable in spite of being conceptuallydifficult. The key is distinguishing betweenthree different issues: the basic definitionof competence, the resources, and thedetails of measurement. While the level ofrequired competence is conceptually andpolitically difficult, society does regularlyresolve it. Many political and socialprocesses operate to define minimumfunctional skills in society, including theofficial acts of legislative bodies in settingstandards and the demands of the businesscommunity for skilled workers. Standardsetting for education is, in fact, both funda-mental and historically well established.The gradual spread of mass educationthrough successively higher levels ofschooling, culminating in universal highschool education, is one historical example(Chambers and Parrish 1983). Providinghigh school education for all those whowould otherwise not obtain it is consider-ably more expensive than the incrementalcost of raising minimum scores, whateverthat cost might be. Indeed, public education

any necessary affordable amounts of addi-tional funding will probably be forthcom-ing. More importantly, the proper directionfor educational reform will be clarified as welearn about productive modifications ofexisting systems of educational governanceand finance.

To show the importance of research, thispaper must demonstrate both the promiseand incompleteness of the existing knowl-edge base. The organization of the paperwill outline the empirical “argument” foradequacy, including the building blocks of apolicy-relevant agenda, showing within eachelement the importance of the topic, theproblem for research, and the suggestedagenda for further research. Accordingly,each section which follows addresses adistinct area of research important to a betterunderstanding of the issue of educationaladequacy.

The definition and level ofminimum educationaloutcomes

As if the problems of actually achievingit were not difficult enough, adequacy alsoconfronts a number of serious, thresholdproblems of definition and political accept-ability (Chambers and Parrish 1983). “Ad-equate” means sufficient for some outcome.However, the definition of the target out-comes and the availability of possible meansto the outcomes are both socially controver-sial and subject to competing demands forother private and social goods. Whatpackage of social outcomes, such as eco-nomic productivity and good citizenship, isto be produced by the schools? And, at whatlevels are these outcomes to be achieved—through what kind, duration, and intensity ofschooling, and at what permissible cost?Further compounding the complexity is thequestion of what value to place on equality.Societies have aggregate “taste” for equality(or distaste for inequality, or at least human

...adequacy alsoconfronts anumber of

serious, thresholdproblems of

definition and

politicalacceptability

(Chambers and

Parrish 1983).


itself is based on a notion of adequacy. Thecombination of courts recognizing adequacyas one of the purposes of public education,and state legislatures defining minimumfunctional levels of achievement, is prob-ably the main example of how social pro-cesses do, in fact, resolve the indeterminateconceptual difficulties. It is noteworthy thatmainstream institutions have a strongincentive for accuracy in defining adequacyand no obvious political bias. Commercialenterprises, for example, need workers witha minimum of general skills, but do not wantto pay for more skills than are necessary.

The second question about adequacy isthat of available resources. Here, one mightthink there is an obvious political problem,perhaps even an impasse, over redistributionof wealth. Wealthier taxpayers resisttransfer payments to the poor (Goertz 1983),and protection of minority interests blockedin the legislative process is one of the classicfunctions of courts (Clune 1984). Yetschool finance reform has a much broaderbase than that of some minority interests.Public finance suggests that it is rational forsociety to pay for the external benefits ofeducation, and the political process seems tobehave in approximately that fashion.Educational spending has grown faster thanthe cost of living for many decades (Oddenand Clune 1995a) and maintains its relativeposition even in the midst of calls forausterity. All educational spending involvesa transfer from taxpayers to those beingeducated. There are very few who call forthe outright repeal of educationalsubsidies or even the rollback of universalhigh school education. Compensatory aidfor the poor is enacted by legislatures, evenin the absence of court mandates. Litigationover school finance frequently involvescourts siding with a substantial minority ofaggrieved districts (Alexander 1991;Grossman 1995, forthcoming). It seemslikely then, that a substantial amount ofextra money could be found for adequacy,even ignoring the fact that adequacy pro-

duces substantial savings such as reductionsin retentions in grade and special education(Barnett 1995), and could be financed byrelatively small adjustments in existingfunding. Thus, money probably can befound to support adequacy, if the educationalsystem can be persuasive in removing doubtsabout the weak link between increasedspending and better outcomes (the very topicof this paper).

The third set of issues about definingadequacy concerns how to measure it.Courts frequently develop long lists ofdesirable outcomes for education (Rose v.Council for Better Education, Inc., 790 s.w.Ken. 2d 186, 212-13 [1989]; Underwood1995, forthcoming). A lively debate pres-ently exists about the importance of “basicskills” versus “higher order thinking” fordisadvantaged children (Commission onChapter 1 1992). Recently developed state-of-the-art tests in Kentucky and Californiahave resulted in many students failing toachieve a “proficient” score (Kirst et al.1995). The impression left by such debatesis that the precise operational definition ofadequacy is hopelessly confused. Inevitablythere are many complex issues to be re-solved in actually developing a test ofstudent achievement (Trimble and Forsaith1995, forthcoming).

I am skeptical about the extent of thereal disagreement, however. As withdefining adequacy and resources, there isprobably more theoretical than practicalconfusion. A simple working definition ofadequacy includes basic proficiency inliteracy, numeracy, and problem solving;completion of a standard high school educa-tion; and perhaps eligibility for highereducation. The question, then, is how muchany specific measurement of adequacy reallydeparts from this conceptual core. A reason-able hypothesis is that various tests arehighly correlated, various standards defineapproximately the same level of perfor-mance, and all standards assume continued

Educationalspending hasgrown faster than

the cost of livingfor many decades(Odden and

Clune 1995a)and maintains itsrelative position

even in the midstof calls forausterity.


attendance and progress through the years ofschooling. For example, “high minimumachievement” probably is very close to adifferent standard sometimes recommended:average achievement among disadvantagedchildren equal to the average of the popula-tion as a whole. The higher levels of profi-ciency demanded by avant-garde testsprobably should be regarded as an effort tochange educational content and raise stan-dards for the future.

Research which examines the degree ofconformity among various standards ofadequacy in the United States might helpestablish the existence of social consensusover adequacy and dispel confusion overspecific standards. Clearly, policymakerscould benefit from a better perspective onthe differences and similarities amongvarious operational definitions. It would beuseful to know the fiscal status of poorchildren in terms of availability of resources.How much compensatory aid is alreadyavailable from what sources? How do suchchildren stand relative to others in the stateand in the country when this aid is added tothe base funding of the schools attended bypoor children? Such data could begin toprovide a view of the spending gap foradequate compensatory funding. Fiscal datacould then be added to the data base on highpoverty schools as suggested below.

The extent and distribution ofinadequate educationaloutcomes and the overlappingpopulations of special needschildren

Part of the case for previous reformmovements aimed at meeting special studentneeds was an empirical demonstration of theextent and distribution of the exclusion fromeducational opportunity (Clune and Van Pelt

1985). Available data on high-povertyschools suggest a remarkable picture of lowoutcomes, inadequate inputs, and overlapwith minority status (Berne 1994a; Natriello

1994). The empirical question is whether,for practical purposes, the problem ofeducational adequacy is a problem of high-poverty schools. Existing data suggest thatperhaps 90 percent of the aggregate sub-minimal outcomes in the United States arelocated in perhaps 10 percent of the schools.This is the familiar issue of concentrationvs. diffusion under Chapter 1—the greatmajority of elementary schools get someChapter 1 money (American Institutes forResearch 1993). Further, even includingcompensatory aid, such schools are oftenfunded at or below the average for the state,and have many obvious deficiencies instandard inputs, especially lack of a quali-fied and stable teaching staff (Berne 1994b;Kozol 1991). But, as with the earlier reformmovements, existing data have been devel-oped for litigation or other special purposesand are available only for certain states andcities. Development of a national databasecomparing high-poverty schools to otherschools on multiple dimensions would be animportant contribution.

A second important issue is the extentof overlap of various categories of specialneeds children within the disadvantagedcategory. On one hand, linguistic limita-tions and handicapping conditions do notrepresent the same need as poverty, and theconditions may be cumulative. On the otherhand, there is some overlap. Nominallydifferent programs may be triggered by thesame set of educational indicators (e.g., lowtest scores); and the same remedy may beappropriate for different conditions (seeMadden, Slavin, Karweit, Dolan and Wasik[1991] on using special education teachersfor acceleration of reading for all students).The issue of overlap has both financial andpedagogical consequences—the amount ofaid required and the kind of appropriateeducational interventions.

The sheer feasibility ofsubstantially raisingminimum performance

The empiricalquestion is

whether, for

practicalpurposes, the

problem of

educationaladequacy is a

problem of high-

poverty schools.


Much confusion still exists about

achievement of disadvantaged children up tohigh minimum performance, within any

The BellCurvethe position that remedial education pro-duces small, uncertain gains with ordinary

by high SES families produces a big effect.Ralph and colleagues maintain that the

stable and little influenced by quality ofschooling. For example, children in the

up to the 4th grade performance of theirhigher achieving classmates until practically

Crouse 1994). On the other hand, Slavin

raise the reading achievement of elementarystudents to grade level (average) “whenever

here is alsosome evidence that the lower achievement

represents nothing more than the accumula-tion of small annual increments of a magni-

schooling (Gamoran in press). Additionally,some research on production functions and

feasible increases in resources (Clune1995b; Duncombe, Ruggeriero, and Yinger

One persistent source of confusion insuch research is the difference between

rion references. Even a small yearly gain instandard deviations or percentiles may

high minimum performance. Failure tofocus on lower achieving students is another

which examines the data on disadvantagedchildren in terms of criterion-referenced

clarify the picture.

Another obviously important source of

(in the generic sense which includes allschools with an accelerative mission). theoretical ideal would be to have acceler-ated schools cooperate in some kind of

could be learned from readily available datawhich is simply not reported. Accelerated

budget and often lack the resources neededfor a thorough evaluation. Slavin’s “success

exception in providing longitudinal data onstudent achievement. Even these schools,

data on costs (Barnett 1995). While feasibil-ity nal achievement data, cost is also of greatimportance because of the political relevance

Educational andimplementation processes are additional

should be investigated under the generalheading of feasibility, addressed below.

evaluation data base for a group of acceler-ated schools would seem to be well worth

themselves are already expending muchhigher levels of resources necessary to

Geographic mobility, irregularattendance and the problem ofpartial and sharedresponsibility among schools

Geographic mobility of students attend-ing urban schools is a problem for adequacyin several ways. It is a serious challenge tothe model of school accountability andschool improvement because schools cannotbe held responsible for the full annualachievement of students who spend a shortertime in the school (Ferguson and Ladd1995). Behind accountability lie problems ofenrollment management and organization ofservice delivery. Student mobility can be

...Slavin and hiscolleagues (1994)claim to be able

to raise thereadingachievement of

elementarystudents to gradelevel (average)

“whenever andwherever wechoose.”


reduced to an unknown extent by administra-tion (e.g., altering enrollment zones andproviding transportation) and by encourag-ing stability through the quality of educationat particular schools. In the latter case,mobility is considered an “endogenous”problem of school quality (Barnett 1995).To the extent that changes in attendancecannot or should not be avoided, the ques-tion is how to organize the delivery ofeducational services (e.g., how to coordinateprograms under conditions of shared respon-sibility). Mobility seems to affect studentachievement, that of both the transferringstudents and the other students (Fergusonand Ladd 1995). The delivery system,however, is fundamentally blind to mobilityand organizing instruction based on anassumption of continuous instruction in oneplace. Mobility also may have stronglydifferential effects on different kinds of highpoverty schools, creating a system of un-equal, stratified responsibility in urbanschool systems. Magnet schools, for ex-ample, not only attract the most qualifiedstudents, but are able to control attendancein advance through the process of selection.Neighborhood schools, on the other hand,are stripped of the most able students andmust bear the brunt of all of the problems ofstudent mobility, including late registrationand high annual turnover (Moore 1990).Irregular attendance is both an additionaland compounding factor for schools’ respon-sibility.

Because mobility may be an important,unmeasured aspect of educational need, it isimportant that we obtain more and betterdata on the extent and kind of mobility andattendance in the variety of high povertyschools. Where possible, the effect ofmobility on achievement should also beanalyzed.

The educational technologiesof accelerated education

Accelerated education is the ultimategoal of educational adequacy (Levin 1988).If poor children could learn the samematerial as other children at the same pace,they would not be educationally disadvan-taged. A longitudinal study of students inChicago found that only four percentgraduated with a reading level above thenational average. Two key educationaloutcomes where the students lagged behindwere devastatingly significant in relation toacceleration: graduation from high school,and reading as a vital measure of studentachievement (Moore 1990).

In one sense, we have known the basicelements of successful accelerated educationfor some time: an accelerated curriculum(or, more accurately, not a retarded curricu-lum), high expectations for student learning,a positive school climate (for both studentsand staff), and a safe and orderly environ-ment (Purkey and Smith 1983). In anothersense, however, these are intermediateoutcomes, or indicators, of a successfuleducational process, which are not self-explanatory about how such desirablecharacteristics are produced. It is obviousthat an accelerated curriculum is a key partof accelerated education, but it is extremelyunlikely that accelerated education can beproduced just by changing the content ofcourses and textbooks. Appropriatelytailored subject matter, skills of the teachers,and the minds of the students must convergearound a complete learning process which israre in high poverty schools.

The central question, then, is what thestaff of the school must do that is different.Research has also established variousanswers to this question: staff developmentto raise expectations and acquire newteaching skills, extra staff time spent assist-ing students with learning problems, andoutreach to the families and caretakers. (SeeBarnett [1995] for an overview of threeapproaches to accelerated education.) Thekey gap in our knowledge concerns how

...we have knownthe basic

elements of

successfulaccelerated

education for

some time: anaccelerated

curriculum...,

high expectationsfor studentlearning, a

positive schoolclimate..., and asafe and orderly

environment(Purkey and

Smith 1983).


these elements can be combined, the condi-tions under which various combinations aresuccessful, and, ultimately, the feasibilityand cost of various approaches. For ex-ample, if the staff development throughoutside consultants, alone, is enough for thetypical high poverty school, the cost wouldbe relatively minor (Barnett 1995). But ifthat model is successful in a small percent-age of schools, additional, more expensivechanges may be also required in the typicalschool. These changes might includepreschool, extra staff for tutoring, highersalaries for more skilled teachers, and ahigher budget for school safety and security.Even the best designed program, supportedat the highest resource level, may fail whereconditions are most unfavorable. As aminor example, high salaries will not attractand maintain qualified teachers if personnelpolicies are unacceptable (Murnane et al.1991).

Knowledge combining educationalpractice, cost, outcomes, and feasibility isdifficult to obtain, but greater success ispossible. The main challenge is that condi-tions of success in a limited number ofschools may be difficult to measure andreplicate. For example, low cost modelsmay flourish in schools where there areunusual leaders and teachers, and initialsuccess may have a powerful impact overtime on the quality of both staff and stu-dents. Since we cannot know what workson a broad scale until we try it, the mostlogical approach would seem to be a pro-gram of phased experimentation in whichvarious programs are evaluated with variouslevels of resources (Clune 1994a, 1995b). Itseems strange that a society so preoccupiedwith the importance of educational ad-equacy and cost should spend so little oneducational experiments, which are rela-tively inexpensive compared to any kind ofbroad scale increase in resources. Perhapswe are only now reaching the point wherethe central questions are becoming clear.

Short of a program of organized experi-ments, we still could gather much better dataabout current interventions. This paper hasalready recommended that better data beobtained on the outcomes and cost ofsuccessful accelerated schools. This sectionadds the element of educational process, orteaching technology. We need a fine-grainedunderstanding of what staff and students dodifferently in order to begin understandingthe conditions under which such desirablebehaviors can flourish. Careful investiga-tions of schools adopting models of acceler-ated education are, therefore, a promisingdirection for research.

The extra costs of acceleratededucation

The extra costs of accelerated educationare currently the subject of a wide diver-gence of views. A high estimate was madein Abbott v. Burke (119 N.J. 287; 575 A. 2d359, 400 [1990]), the New Jersey case whichordered spending in special needs districtsbe set equal to resource-rich suburbandistricts, with an added supplement forremedial education. The New Jersey Courtreasoned that a good estimate of costs ofadequacy is the resources available to thestate’s most successful students, e.g., thosewith the best outcomes. An estimate of thesame magnitude was derived through costanalysis by Duncombe, Ruggeriero, andYinger (1995) resulting in at least an addi-tional $3,800 per pupil in New York State.A low estimate was derived by Barnett whosynthesized cost-effectiveness research onthree programs of accelerated education andconcluded that the extra costs amounted tobetween $100 and $1,000 per pupil per year,an amount easily covered by existing rev-enues (Barnett 1995). Barnett also con-cludes, however, there is no data on the basefunding of these programs, their successrate, or the extent to which the interventionscan be generalized. The most reasonableexplanation of Barnett’s results is thesuccessful interventions represent the low-

The New JerseyCourt reasonedthat a good

estimate of costsof adequacy isthe resources

available to thestate’s mostsuccessful

students, e.g.,those with thebest outcomes.


cost, high-skill end of a production frontier(Clune 1995b), achieved in certain schoolsand not in others by reformers who assumethey must operate within existing levels ofresources (Clune 1995b). For this reason, Idid not modify my earlier guess of the fullrange of costs—between $2,000 and $5,000per pupil per year. On the other hand, wecannot reach any firm conclusions withoutmore data, and there is certainly no reason topay more than is necessary.

A large part of the problem in estimatingcosts is the failure to specify the variouscomponents of costs and gather appropriatedata. Adequacy-based funding is thought toinvolve three major components: basefunding, student needs, and extra costs forstandard inputs (Chambers and Parrish1983). The discussion begins with studentneeds, another name for accelerated educa-tion, but add the element of “slack” in theexisting system. The extra costs of acceler-ated instruction actually consist of the totalcosts minus any savings which can begarnered from the typical preexisting educa-tional program. As discussed in the previoussection, the costs of accelerated educationdepend on the combination of inputs effec-tive in various contexts. Professionaldevelopment alone tends to be inexpensive.Extra staff, higher pay for teachers, and newlevels of education such as preschool aremore expensive (but also seemingly afford-able).

The next question is how much “slack”is available in existing schools to pay foradditional costs. Under one view, there isconsiderable slack, for example, in thefollowing: high expectations and a challeng-ing curriculum replacing low expectationsfor learning and dumbed-down curriculum;coordinated vs. uncoordinated learning;active time vs. down time; and efficient vs.inefficient instructional methods. These areall possible with relatively minor expensesfor professional development. There is alsoa question about wasted administrative time.

Under Levin’s model, all faculty meetingsare used for planning and implementingaccelerated education. This seems toassume that large parts of previous facultymeetings were unnecessary. Anotherpossible area of slack is the conversion ofexisting positions. One study found thatextra positions could be made available byefficient management of class size (Miles1994). Some accelerated schools appear toput more teachers in the classroom (Darling-Hammond 1995, forthcoming). Finally,accelerated schools may consolidate pro-grams and positions aimed at meeting avariety of special needs. Slavin’s schoolsvirtually eliminate special education andretention in grade, focusing almost allremedial resources on catching up inreading, mathematics, etc. (Barnett 1995).Generally, claims of available slack aremade somewhat more plausible becauseprograms which focus on specific learninggoals have incentives and standards formarshalling resources which are lacking inorganizations with multiple, fragmentedgoals. On the other hand, the opposite viewof a narrow focus is the possibility thatother unmeasured goals have been sacri-ficed.

What is the impact of Slavin’s programon the achievement of handicapped studentswho previously may have received moreinstructional resources (Levin 1994)?Teachers commonly claim that almost anystrongly evaluated program with a singlefocus will cause them to sacrifice importantparts of their teaching. For example,evaluation based on mathematics scoresmay overlook the importance of fine arts ina particular community. The possibledistorting effects of accountability are avexing issue, but the risk may be tolerableas long as in the aggregate and in importantsub-groups students perform well on a broadrange of skills considered most important bysociety. One of the most enduring findingsof education research is that programs scorehigher against measures of goals at which

Adequacy-basedfunding isthought to

involve threemajor

components:

base funding,student needs,

and extra costs

for standardinputs

(Chambers and

Parrish 1983).


they are aimed than on generally applicabletests (Walker and Shaffarszick 1974). Thisfinding seems better interpreted as anendorsement of organizational focus than asan assumption that all organizations mustoptimize some set of measured and unmea-sured goals.

The assumed base level of spending is asomewhat confusing and complex issueunder an adequacy-based approach calcu-lated in terms of incremental costs. Assum-ing that extra costs are $2,000 per pupil, thequestion is: $2,000 per pupil added towhat? The normal starting place is horizon-tal equity, an assumed base of spendingguaranteed for every district and student inthe state in the form of spending per pupil orin educational vouchers. Conceptually, thisbase level of funding should correspond toan adequately financed program for a schoolwhich has no extra costs from expensiveinputs or special needs—in effect, it is theleast-cost school in the state (Duncombe,Ruggeriero, and Yinger 1995). Odden andClune (1995b) suggest the 90th percentile ofrural spending as a rough approximation ofsuch a school, but there are alternativemethods. The Alabama litigation, forexample, used various official minimumstandards for inputs to establish the cost of abase program (Hershkoff, Cohen, andMorgan 1995). Guaranteeing a reasonablebase to high-poverty schools is much moreimportant than the exact method of calcula-tion. Many high poverty schools are locatedin resource-constrained school districts.This means that any available compensatoryaid must be used to support the basicprogram rather than provide acceleratededucation.

The final component of adequacyfinancing is extra costs for standard inputs(Chambers and Parrish 1983). High povertyschools seem to be burdened with five typesof high input costs: preschool and Kinder-garten, qualified teacher salaries, buildingand maintenance, school safety, and social

services. Early childhood education isassumed to be especially important andbeneficial for disadvantaged children(Barnett 1995); and Slavin’s programsguarantee some amount of this input (Mad-den et al. 1991). Teacher salaries must behigher in less desirable educational settings(Chambers 1981; National EducationAssociation 1995; Reschovsky and Wiseman1995). As for capital expenses, a commonfinding of the implementation of schoolfinance reform is that poor schools fre-quently draw on instructional revenues topay for deferred costs of building andmaintenance (Firestone et al. 1994; Picusand Hertert 1993). School safety is widelyconsidered an important aspect of effectiveschools (Purkey and Smith 1983), but less isknown about how to estimate costs. Socialservices are the most confusing of thecategories because they are potentiallyimportant in the case of poor children andmay be organized separately (despite themovement toward collaboration and school-linked services, and are financed by differentbudgets [Kirst 1994; Zigler and Stevenson1994]).

One of the most confusing issues inadequacy-based funding is how to measureand do research on these various categoriesof costs. Cost analysis attempts to reflect thebroadest range of costs, including teachersalaries. However, the sample used may notbe representative of the most needy schools,and capital expenditures usually are notincluded (Clune 1995b; Duncombe,Ruggeriero, and Yinger 1995). At the otherextreme, cost-effectiveness analysis ofaccelerated programs often does not takeinto account either the base program orcostly inputs. These are both large catego-ries relative to extra instructional spendingdesigned to meet student needs. It would,therefore, be helpful if future research couldbegin to triangulate cost estimates drawnfrom different assumptions and includeddifferent components of cost. Studies ofaccelerated schools could gather better data

One of the mostconfusing issuesin adequacy-

based funding ishow to measureand do research

on...variouscategories ofcosts.


on base funding and input costs in additionto the marginal cost of meeting studentneeds. Cost analysis might focus on high-poverty schools and include capital spend-ing. Research on labor markets for teachersmight attempt to calibrate costs relative toqualifications shown by research to beimportant to the achievement of disadvan-taged students. Capital costs may benefitfrom a separate analysis based on rationalstandards of a physical plant which isminimally acceptable for adequate instruc-tion of the type required for acceleratededucation. Such analysis of capital costsshould take into consideration any backlogof needs and the effects of changes instudent population (e.g., rapid growth).

The structure of state aid underan adequacy theory

The basic flaws of existing educationalfunding formulas, from the perspective ofeducational adequacy, are quite simple andtwofold: failure to provide either guaranteedbase funding or fully cover extra costs ofinputs and special needs. At a deeper level,these flaws are caused by starkly differentassumptions of adequacy and equity ap-proaches to state aid (Reschovsky andWiseman 1995). Most state aid formulas arebuilt around inter-district equity, spend moremoney on horizontal equity than specialneeds or extra costs, and include a generouselement of local taxing discretion. Theseformulas may or may not be supported by afair system of matching grants from the state(guaranteed tax base). Previously, I haverecommended that the special needs part ofthe formula be fully funded and put on top ofa strong statewide system of horizontalequity. For example, one might think of afoundation level at the 90th percentile ofrural spending, a guaranteed tax base abovethat level, categorical aid for instructionalneeds, and special adjustments for extracosts (Clune 1995a, forthcoming).

Such an approach of supplementedequity would certainly represent an im-provement for many high poverty schools(see American Institutes for Research 1993:how the lowest levels of compensatory aidunder Chapter 1 often went to the poorestdistricts). However, there remain someserious problems. The generous guaranteeof horizontal equity in the form of the highfoundation program causes competition forscarce funds between horizontal equity andspecial needs. Local taxing discretion islikely to result in inadequately addressingthe special needs of some of the state’sneediest students. An alternative is toforthrightly recognize the different assump-tions of the two systems and design aseparate system of finance for high povertyschools (similar to what the court attemptedin New Jersey but with more refinedestimates). High-poverty schools couldreceive a high foundation grant, specialcategorical aid, and supplements for extracosts, all guaranteed by the state under areasonable system of sharing state and localtaxes. Such a distinct and severable ap-proach to the funding of high povertyschools probably also represents a better fitwith the special problems. These problemsof governance and accountability, discussedbelow, arise from the attempt to createbetter incentives for actually adopting someform of accelerated education.

What kind of research is possible onschool aid formulas? The most fruitfulapproach appears to be a set of simulationsbased on different assumptions of the costsof adequacy and the structure of state aid.The basic questions are the same for everystudy of school finance: total revenue, taxburden, and distributional consequences(how much would it cost, to which taxpay-ers, and how many schools would experi-ence an increase or decrease in spending).While many simulations have been done ofschool aid formulas (Odden and Picus1992), a different view would be producedby a rigorous analysis focused on the special

The basic flawsof existing

educational

[adequacy]funding

formulas...are...

failure to provideeither guaranteed

base funding or

fully cover extracosts of inputs

and special

needs.


needs of poor children. The question wouldbe one of “opportunity cost” (Chambers andParrish 1983)—what are the revenue anddistributional consequences of full fundingof adequate education while meeting theother needs in a cost-minimizing fashion.Of course, the normal assumption is exactlythe opposite—meeting various competingneeds and determining what is left over forthe poor; and this tends to hold true even inthe context of litigation aimed primarily atcompensatory education (Goertz 1983). Insuch studies, it would be difficult to ignorethe problem often discussed in schoolfinance of inter-state differences in spend-ing. Alabama clearly has lower costs thanNew York, but such costs are not propor-tionate to the full difference in wealth andspending. A suitable national policy onadequate funding (what Chapter 1 purportsto be) would need to take into considerationinter-state differences in the level of basefunding. Therefore, research on adequacyfunding should include simulations of theeffects of interstate adjustments.

Implementation—increasingthe number of schools usingcost-effective acceleratededucation

Suppose, then, we understand whataccelerated education is, how much it willcost, and how funding is obtained. Howdoes the policy system guarantee, or in-crease the odds, that the schools receivingthe funds actually will adopt and success-fully implement a version of acceleratededucation? Whatever the educationaltechnology and cost, accelerated educationinvolves adopting a set of new attitudes,teaching skills, and management practices.On one level, the greatest puzzle aboutaccelerated education is why it is adoptedby so few schools. Does the barrier towider implementation consist of motivationor capacity, or both? A serious debate aboutthis exists, somewhat paralleling the debateover costs. At one end of the spectrum, the

transfer is seen as relatively easy, and can beimplemented on a wide scale by a combina-tion of state accountability, incentives, andprofessional development (Hanushek 1994).An intermediate position on implementationdifficulty is that schools will need to joinaccelerated schools’ networks and receivethe intensive, targeted training availablethrough those networks (Clune 1987; Slavin,Madden, and Dolan 1994). A third view isthat we have barely touched the problem ofimplementation, because existing successfulschools are probably connoisseurs of im-provement (Elmore 1995, forthcoming).Most schools, under this view, presentserious problems of inertia and resistance towholesale improvement. These are prob-lems which are presently little understood,either in their dimensions or remedies.

Lack of data is one problem in answer-ing the questions about implementation.Implementation of accelerated education iseven less studied than evaluation of costsand outcomes. The establishment andsurvival of demonstrations and experimentsis an understood part of published literature.Little thought is directed to the problematicnature of how such innovations are dissemi-nated, adopted, and transferred. Morestudies of propagation, adoption, andtraining within accelerated school networksare needed. Schools which successfullyadopt a change should be compared withthose where the effort fails. Another impor-tant comparison is that of schools understrong pressure and incentives to improvefrom state and local accountability systems(as opposed to networks). Is improvementlargely random under the different systems?Is there a residue of schools resistant to anystimulus, and what is known about how toreform such schools? For example, howdoes the remedy of closing failing schoolswork?

Implementation also involves some costsbeyond professional training in the targetschool. Member schools of the Success for

Lack of data isone problem inanswering the

questions aboutimplementation....More studies of

propagation,adoption, andtraining within

acceleratedschool networksare needed.


All network have been providing a trainedteacher/manager to other schools for a year,free of charge. To the extent that propaga-tion does involve a subsidy, it is necessary tounderstand the best method of financing thiscost for the future.

Governance—the blend ofaccountability variety anddecentralization necessary foradequate education

Whatever the proper stimulus forreform, a question arises about the range ofgovernance structures capable of toleratingand encouraging such reforms. Consider-ation of this range of structures will againresult in serious debate. In one sense, theproblem is the familiar tension between top-down and bottom-up reform. Research onimplementation will probably clarify that notmuch will happen without a stimulus tochange. It seems equally clear, however, thatthere are at least several kinds of acceleratededucation, each requiring a high degree ofadaptation and flexibility. The question,then, is what kind of sensitive governancestructure is capable of exerting strongpressure without stifling schools or pushingthem in the wrong direction.

Some authorities seem to believe thataccelerated education can occur under mostkinds of governance systems. It can beviewed as traditional education done moreeffectively and efficiently and, therefore,presents no problem for management orunions (this is my impression from talkingwith Henry Levin). I took the positionearlier that strong instructional guidancefrom central government was an unlikelymatch (Clune 1987), but arguments for thepower of coherent policy have not beenproven wrong (Smith and O’Day 1991). Thepower of simple incentive systems, such asrewards for outcomes or participation innetworks, deserve consideration because oftheir relatively non-intrusive character(Hanushek 1994). Some people regard

school districts, especially large urbandistricts, as a serious obstacle (Darling-Hammond 1995, forthcoming) while othersthink the relationship varies according tolocal context (Carnoy et al. 1994). Underone view, many large districts are theenemies of school improvement, not onlystifling them with unnecessary regulations,but regarding their success as a threat. Thisview probably favors the largest possibleamount of site management but, addition-ally, does not explain the source of pressureand incentives to encourage schools tochange. If it is necessary that the schoolsjoin a network, what agency will certify thenetwork: what are the incentives; and howshall good faith participation be evaluated?In an earlier article, I suggested that net-works perform technical assistance andprovide simple data for evaluation, leavingonly ultimate evaluation to a governmentagency (Clune 1987). On the basis ofempirical evidence, some researchers areskeptical about the capacity of any govern-ment agency to assess when intervention isrequired and to intervene in an appropriatemanner (Firestone and Nagle 1995; Hess1994). The deregulation and community ofpurpose attributed to radical decentraliza-tion, e.g., choice and charters, have beenrecommended as possible sources of schoolimprovement, or at least of freedom fromburdensome regulations; and these claimsshould be seriously evaluated.

The research which seems most appro-priate to these questions would be studies ofrapidly improving and other schools in bothnetwork and incentive systems undervarious conditions of guidance, restriction,and assistance from central government.

Choice and decentralization—the reality of a low cost, lowchange alternative

In a political sense, choice and decen-tralization seem to have a quicker start thansustained school improvement, perhaps

The question,then, is what

kind of sensitive

governancestructure iscapable of

exerting strongpressure without

stifling schools

or pushing themin the wrong

direction.


because they are easier do. Despite earlierclaims (Chubb 1990; Chubb and Moe1990), the evidence that decentralizationwill produce major changes in achievementseems weak, both empirically and theoreti-cally (Clune 1990; Hoxby 1995; Witte et al.1994). Indeed, the rationale of choiceadvocates may be changing from instrumen-tal arguments about effectiveness to valuearguments about freedom of choice andassociation. But choice does offer at leastone important option in the debate overadequacy: the possibility of a low cost, lowchange alternative. The Milwaukee experi-ment did not increase scores, but it did notdecrease them either, and the voucherschools cost much less than the publicschools. Again, the generalization ofmodest cost is subject to vehement debate.Initially, low cost seems based on excesscapacity in a limited number of existingprivate schools and a secondary labormarket for teachers. It is doubtful thatdrastic reductions in teacher qualificationsresulting from massive cuts in teachersalaries are politically feasible, or will havea beneficial impact on student achievement.

Nevertheless, these are all empiricalquestions, and the political appeal of choicemakes it well worth watching. From theperspective of adequacy, we need to under-stand effect of large scale decentralizationchange on education for poor children. Areasonable set of questions would includethe effects on: private investment; publicinvestment; costs (including transportationand regulation, [Levin 1990]); the labormarket for teachers; student achievement;and stratification of students by race, classand achievement. One problem for suchresearch is finding large scale demonstra-tions in the United States. Choice systemsin this country tend to be small and incre-mental. Comparative studies involving

many students, such as those conducted inEurope and South America, may be useful,especially with regard to some basic dynam-ics of student achievement, markets andpublic finance (Carnoy 1995).

A crosscutting look at theresearch questions andmethodologies appropriate forstudying educational adequacy

This section summarizes researchquestions discussed in this paper and com-ments on methodologies appropriate foranswering them.

1. Definition of educational ad-equacy and the political problem ofavailable resources

• degree of conformity among minimumstandards of achievement in the UnitedStates

• existing funding levels for poor studentsand the extent of the funding gap

2. Demographics of educational need

• distribution of low educationaloutcomes in the U.S. and degree of correla-tion with high-poverty schools

• overlap and divergence of variouscategories of special needs: poor, handi-capped, bilingual

3. Feasibility of raising educationaloutcomes

• studies of data on poor children

• longitudinal outcome data fromaccelerated schools

Comparativestudies involvingmany students,

such as thoseconducted inEurope and

South America,may be useful,especially with

regard to somebasic dynamicsof student

achievement,markets andpublic finance

(Carnoy 1995).


4. Geographic mobility and irregularattendance

• studies of the extent and kind ofmobility in high poverty schools and itsimpact on student achievement

5. Educational technologies ofaccelerated education

• organized experiments of acceler-ated education involving different programapproaches and costs

• studies of the teaching technologiesof existing accelerated schools

6. Extra costs of accelerated educa-tion

• parallel cost estimates from acceler-ated schools and quantitative analysis

• separate estimates of capital costneeds

• studies of the costs of qualifiedteachers in the labor market

7. Structure of state aid under anadequacy theory

• simulations of two different aidformulas under different cost assumptions

8. Implementation—increasing thenumber of schools using cost-effectiveaccelerated education

• studies of attempts to implementnew accelerated schools

• studies of school transformationsunder various incentive systems

• studies of the costs of school im-provement networks

9. Governance—the blend of ac-countability variety and decentraliza-tion necessary for adequate education

• studies of role of governance inrapidly improving and other schools underdifferent kinds and degrees of educationalgovernance (including decentralization)

10. Educational decentralization as alow cost alternative

• studies of investment, cost andeducational outcomes especially in largescale choice experiments

Several of generalizations can be madeabout the research methods appropriate tothese questions. First, much of the agendadepends on gathering more data fromexisting accelerated schools. Acceleratededucation is not very common in U.S.education, so it is not surprising that thesome of the most valuable data would comefrom demonstration sites. At the low end ofthe scale of research burden, it would bevery helpful to have basic data on educa-tional outcomes and costs. At a higher levelof burden, we would like a much betterunderstanding of teaching technology,implementation, and fine-grained aspects ofcosts. But obtaining systematic data fromaccelerated schools presents a problem.The schools have their own systems ofevaluation and may not agree with exter-nally developed standards. Data gatheringis also intrusive and expensive. Thissuggests a rather intense need for some kindof cooperative agreement and, if possible,extra financing for research.

A second category of research involvesa variety of potentially useful secondarydata sources—the demographics of studentneed, achievement trends among the poor,geographical mobility, simulations offunding formulas, and the like. Someresearch has been done in all of these areasand will continue to be done. The problems

The problems forfashioning a truenational agenda

involvesupporting andcoordinating a

more extensiveand cohesive set

of related studies.


for fashioning a true national agendainvolve supporting and coordinating a moreextensive and cohesive set of relatedstudies. Such an effort might normally beorganized by the government or a founda-tion; but some real progress might bepossible through the voluntary cooperationof research scholars and interested govern-ment agencies, such as big city schooldistricts.

A third category of research involves astudy of change in schools operating underdifferent systems of incentives and gover-nance. This kind of research seems bestsuited to a large scale, longitudinal researchproject, such as a research center.

The fourth category of research in-volves implementation of large systems ofeducational choice. Such research also mayneed substantial funding but may occur inselected locations without any special effortbecause such systems are rarely imple-mented and are the subject of great schol-arly and political interest. Input from theadequacy community into the design of theresearch might be valuable.

Conclusion: The cumulativeeffect of better information

The basic thesis of this paper has beenthat a comprehensive research effort canhelp establish the legitimacy of a programof educational adequacy and define itscontours. Much fragmented research hasbeen done on all of the research topicsdiscussed. It is hoped that an orderlyexposition of the full range of topics willgenerate enthusiasm for an effort which hasa higher level of support and better coordi-nation.

...a

comprehensiveresearch effortcan help

establish thelegitimacy of aprogram of

educationaladequacy...


References

Alexander, K. 1991. “The Common School Ideal and the Limits of Legislative Authority: The Kentucky Case.”Harvard Journal on Legislation 28(2): 341.

American Institutes for Research. 1993. Translating Dollars into Services: Chapter 1 Resources in the Context ofState and Local Resources for Education. Palo Alto, CA.

Barnett, S. 1996. “Economics of School Reform: What Can We Learn from Three Promising Models.” in HoldingSchools Accountable: Performance-based Reform in Education. Ed. H.F. Ladd. Washington, DC: The BrookingsInstitution. pp. 299-326.

Berne, R. 1994a. “Educational Input and Outcome Inequities in New York State.” Outcome Equity in Education.Ed. R. Berne and L. Picus. Thousand Oaks, CA: Corwin Press.

Berne, R. 1994b. “Program Equity for All Children: Meaningful Goal or Impossible Dream?” Educational Policy8(4): 502-517.

Carnoy, M. July, 1995. “Is School Privatization the Answer? Data from the Experience of Other Countries SuggestNot.” Education Week XIV(40): 52, 40.

Carnoy, M., J. Hannaway, M. Chun, S.J. Stein, and P.L. Wong. 1994. “Urban and Suburban School Districts: AreThey Different Kinds of Educational Producers.” Paper prepared for the Consortium for Policy Research in Educa-tion Policy Center, Rutgers University, New Brunswick, NJ, and presented at the American Educational ResearchAssociation annual conference.

Chambers, J.G. 1981. “The Hedonic Wage Technique as a Tool for Estimating the Costs of School Personnel: ATheoretical Exposition with Implications for Empirical Analysis.” Journal of Education Finance 6(3): 330-354.

Chambers, J. G., and T.B. Parrish. March, 1983. Adequacy and Equity in State School Finance and Planning: AResource Cost Model Approach. Stanford, CA: Institute for Research on Educational Finance and Governance,Stanford University.

Chubb, J.E. 1990. “Political Institutions and School Organization.” Choice and Control in American Education.1: 227-234. Ed. W.H. Clune and J.F. Witte. Philadelphia: Falmer Press.

Chubb, J.E., and T.M. Moe. 1990. Politics, Markets, and America’s Schools. Washington, DC: The BrookingsInstitution.

Clune, W.H. 1984. “Courts and Legislatures as Arbitrators of Social Change (review of ‘Educational Policymakingand the Courts: An Empirical Study of Judicial Activism’).” Yale Law Journal 93(4): 763.

Clune, W.H. 1987. “Institutional Choice as a Theoretical Framework for Research on Educational Policy.” Educa-tional Evaluation and Policy Analysis 9(2): 117-132.

Clune, W.H. 1990. Introduction to Choice and Control in American Education. 2: 1-15. Ed. W.H. Clune and J.F.Witte. Philadelphia: Falmer Press.


Clune, W.H. 1994a. Introduction to “The Cost and Management of Program Adequacy: An Emerging Issue inEducational Policy and Finance.” Educational Policy 8(4): 365-375.

Clune, W.H. 1994b. The Shift from Equity to Adequacy in School Finance. Educational Policy 8(4), 376-394.

Clune, W.H. 1995. “Accelerated Education as a Remedy for High Poverty Schools.” University of Michigan Journalof Law Reform 28(3): 655-680.

Clune, W.H. April, 1996. “Comments on Chapters Eight, Nine, and Ten.” in Holding Schools Accountable: Perfor-mance-based Reform in Education. Ed. H.F. Ladd. Washington, DC: The Brookings Institution. pp. 357-367.

Clune, W.H. 1996. “Education Finance in the 21st Century: Adequacy and Efficiency.” in Linking EducationFinance and Performance: Selected Proceedings State Education Workshop. Washington, DC: National EducationAssociation. pp. 53-55.

Clune, W.H., and M.H. Van Pelt. Summer, 1985. “A Political Method of Evaluating the Education for All Handi-capped Children Act of 1975 and the Several Gaps of Gap Analysis.” Law and Contemporary Problems 48(1): 7-62.

Commission on Chapter 1. 1992. Making Schools Work for Children in Poverty. Washington, DC: AmericanAssociation for Higher Education.

Darling-Hammond, L. 1996. “Restructuring Schools for High Performance.” in Rewards and Reform: CreatingEducational Incentives That Work. Ed. S. Fuhrman and J. O’Day. San Francisco: Jossey-Bass. pp. 144-192.

Duncombe, W., J. Ruggeriero, and J. Yinger. 1996. “Alternative Approaches to Measuring the Cost of Education.” inHolding Schools Accountable: Performance-based Reform in Education. Ed. H.F. Ladd. Washington, DC: TheBrookings Institution. pp. 327-356.

Elmore, R.F. 1996. “Getting to Scale With Successful Educational Practices.” in Rewards and Reform: CreatingEducational Incentives That Work. Ed. S. Fuhrman and J. O’Day. San Francisco: Jossey-Bass. pp. 294-329.

Ferguson, R.F., and H. Ladd. April, 1996. “How and Why Money Matters: An Analysis of Alabama Schools.” inHolding Schools Accountable: Performance-based Reform in Education. Ed. H.F. Ladd. Washington, DC: TheBrookings Institution. pp. 265-298.

Firestone, W.A., M.E. Goertz, B. Nagle, and M.F. Smelkinson. 1994. “Where Did the $800 Million Go? The FirstYear of New Jersey’s Quality Education Act.” Educational Evaluation and Policy Analysis 16(4): 359-373.

Firestone, W.A., and B. Nagle. 1995. “Differential Regulation: Clever Customization or Unequal Interference?”Educational Evaluation and Policy Analysis 17(1): 97-112.

Gamoran, A. 1996. “Effects of Schooling on Children and Families.” in Family-School Links: How do They AffectEducational Outcomes? Ed. A. Booth and J. Dunn. Hillsdale, NJ: Erlbaum. pp. 107-114.

Gamoran, A. 1994. “The Impact of Academic Course Work on Labor Market Outcomes for Youth Who do notAttend College: A Research Review.” Paper prepared for the National Assessment of Vocational Education.


Goertz, Margaret. Spring, 1993. “School Finance Reform in New Jersey: The Saga Continues.” Journal of Educa-tion Finance 18(4): 346-365.

Goertz, M.E. Work on formula simulations.

Gold, Steven D. May, 1995. The Outlook for School Revenue in the Next Five Years (CPRE Research Report SeriesReport no. 34). New Brunswick, NJ: Rutgers University, Consortium for Policy Research in Education.

Grossman, M. 1995. “Oklahoma School Finance Litigation: Shifting From Equity to Adequacy.” University ofMichigan Journal of Law Reform 28(3): 521-557.

Hanushek, E.A. 1991. “When School Finance ‘Reform’ May Not be Good Policy.” Harvard Journal on Legislation28(2): 423-456.

Hanushek, E.A. 1994. “A Jaundiced View of ‘Adequacy’ in School Finance Reform.” Educational Policy 8(4): 60-469.

Herrnstein, R.J., and C. Murray. 1994. The Bell Curve: Intelligence and Class Structure in American Life. p. 421.New York: The Free Press.

Hess, G.A., Jr. 1994. “Adequacy Rather than Equity: A New Solution or a Stalking Horse?” Educational Policy8(4): 44-567.

Hoxby, C.M. 1996. “The Effects of Private School Vouchers on Schools and Students.” in Holding Schools Ac-countable: Performance-Based Reform in Education. Ed. H.F. Ladd. Washington, DC: The Brookings Institution.pp. 177-208.

Kirst, M.W. 1994. “Equity for Children: Linking Education and Children’s Services.” Educational Policy 8(4):583-590.

Kirst, M. et al. April, 1995. Conditions of Education in California: Excerpts from PACE Report on Assessment andAchievement. Network News and Views XIV(7): 29-47.

Kozol, J. 1991. Savage Inequalities: Children in America’s Schools. New York: Crown.

Levin, H.M. September, 1988. Accelerated Schools for At-Risk Students. CPRE Research Report Series no. RR-010.New Brunswick, NJ: Rutgers University, Center for Policy Research in Education.

Levin, H.M. 1990. “The Theory of Choice Applied to Education.” Choice and Control in American Education.Ed. W.H. Clune and J.F. Witte. 1: 247-284. Philadelphia: Falmer Press.

Levin, H. 1991. “The Economics of Justice in Education.” Spheres of Justice in American Schools, 1990 Yearbookof the American Education Finance Association. Ed. D. Verstegen and J. Ward. New York: Harper & Row. pp. 129-145.

Levin, H.M. 1994. “Little Things Mean a Lot.” Educational Policy 8(4): 396-403.

Madden, N.A., R.E. Slavin, N.L. Karweit, L. Dolan, and B.A. Wasik. 1991. “Success for All.” Phi Delta Kappan72(8): 593-599.


Miles, K.H. 1994. “Finding Time for Improving Schools: A Case Study of Boston Public Schools.” Paper presentedat the annual meeting of the American Educational Research Association, New Orleans, LA.

Moore, D. 1990. “Voice and Choice in Chicago.” Choice and Control in American Education. Ed. W.H. Clune andJ.F. Witte. 2: 153-198. Philadelphia: Falmer Press.

Morgan, M.I., A.S. Cohen, and H. Hershkoff. 1995. “Establishing Education Program Inadequacy: The AlabamaExample.” University of Michigan Journal of Law Reform 28(3): 559-598.

Murnane, R.J., J.D. Singer, J.B. Willett, J. Kemple, and R. Olsen. 1991. Who Will Teach? Policies that Matter.Cambridge: Harvard University Press.

National Education Association. 1995. How Education Spending Matters to Economic Development. Washington,DC.

Natriello, G. 1994. “Four Perspectives on the Disparities Between the Educational Resources Available to Studentsin the Hartford Public Schools and Other Connecticut Communities.” Unpublished paper available from author.

Odden, A., and W.H.Clune. 1995a. “Educational Productivity and Organizing for Higher Performance.” Paperprepared for the Consortium for Policy Research in Education, Rutgers University, New Brunswick, NJ. Draft.

Odden, A., and W.H. Clune. 1995b. “School Finance Formulas: Aging Structures of Renovation.” Paper preparedfor the Consortium for Policy Research in Education, Rutgers University, New Brunswick, NJ. Draft.

Odden, A. R., and L.O. Picus. 1992. School Finance: A Policy Perspective. New York: McGraw Hill.

Picus, L.O., and L. Hertert. 1993. “A School Finance Dilemma for Texas: Achieving Equity in a Time of FiscalConstraint.” Texas Researcher 4: 1-27.

Picus, L. O., and L.A. Toenjes. 1994. “Texas School Finance: Assessing the Equity Impact of Multiple Reforms.”Journal of Texas Public Education 2(3): 39-62.

Purkey, S.C., and M.S. Smith. 1983. “Effective Schools: A Review.” The Elementary School Journal 83(4): 427-452.

Ralph, J., D. Keller, and J. Crouse. 1994. “How Effective Are American Schools?” Phi Delta Kappan 76(2): 144-150.

Reschovsky, A., and M. Wiseman. 1995. “The Two-Thirds Initiative: A Reform Opportunity Lost?” OccasionalPaper Series of the Robert M. La Follette Institute of Public Affairs, University of Wisconsin-Madison.

Slavin, R.E., L.J. Dolan, and J.A. Madden. November, 1994. “Scaling Up: Lessons Learned in the Dissemination ofSuccess for All.” Paper prepared for the Center for Research on the Education of Students Placed at Risk, JohnsHopkins University.

Slavin, R.E., N.A. Madden, and L.J. Dolan. 1994. “Whenever and Wherever We Choose: The Replication of‘Success for All.’” Phi Delta Kappan 75(8): 6639-6640.


Smith, M.S., and J. O’Day. 1991. “Systemic School Reform.” The Politics of Curriculum and Testing, 1990 Year-book of the Politics and Education Association. Ed. S. Fuhrman and B. Malen. pp. 233-267. Bristol, PA: FalmerPress.

Trimble, C. Scott, and A. Forsaith. 1995. “Achieving Equity and Excellence in Kentucky Education.” University ofMichigan Journal of Law Reform 28(3): 599-654.

Underwood, J.K. 1995. School Finance Adequacy as Vertical Equity. University of Michigan Journal of LawReform 28(3): 493-520.

Walker, D.F., and J. Shaffarszick. 1974. “Comparing Curricula.” Review of Educational Research 44: 83-111.

Witte, J., et al. 1994. Fourth Year Report: The Milwaukee Parental Choice Program. Madison, WI: Report to theWisconsin State Legislature.

Zigler, E.F., and M. Finn-Stevenson. 1994. “Schools Role in the Provision of Support Services for Children andFamilies: A Critical Aspect of Program Equity.” Educational Policy 8(4): 591-606.

selected papers in school finance 1995 · 2 selected papers in school finance, 1995 introduction...

Documents