DO.CUMENT RESUME ED 026 .971 .24 HE 000 765 By-Davis, Junius A. Applications of the Science of Measurement to Higher Education. Duke Univ., Durham, N.C. Spans Agency-Office of Education (DHEW), Washington, D.C. Bureau ofResearch. Bureau No- BR -6-1722-32 Pub Date Apr 68 Contract 0EC-2 -6-061722-1742 Note- 1 54p. EDRS Price MF-$0.75 HC-$7.80 Descriptors-Academic Aptitude, Academic' Performance, *Admiision Criteria,*College Admission, *Higher Education,IndividualDifferences, *Measurement Instruments, Readiness (Mental), StudentEvaluation, *Testing Part I of the report provides a historical development ofadmissions procedures in US colleges and universities fromthe seventeenth century to the beginningof the twentieth century. During this period student selection practicesdiffered at each institution but were generally based on prescribedstandards of academic readiness. The need for consistency in requirements led to establishmentof the College Entrance Examination Board, which administered standardized testing acrossinstitutions to evaluate student performance (scholastic achievement)and predict grades in college (scholastic aptitude). The Educational TestingService later became the Board's testing agent to build, administer and score examinations,report test results and conduct necessary research. Part II covers the secondhalf of the twentieth century in which measurement emerged as a science,supplementing measures of academic aptitude and high school performance with measuresof other variables such as interests, motivation, leadership, .and otherindividual student differences. Research organizations or teams in university-based centerscurrently utilize measurement science to study problems such asstudent input factors, influential forces within college environments and their impact on students.These efforts could expand to include studie, on the interaction between students andtheir learning environments, teaching proccAures for heterogeneous student bodies,and the improvement of criteria by which students are evaloated. (WM)

rthee. e

U.S. DEPARTMENT OF HEALTH, EDUCATION& WELFARE OFFICE OF EDUCATION

THIS DOCUMENT HAS BEEN REPRODUCEDEXACTLY AS RECEIVED FROM THE POINTS OF VIEW OR OPINIONS PERSON OR ORGANIZATION ORIGINATING IT.

STATED DO NOT NECESSARILY REPRESENTOFFICIAL OFFICE OF EVCATION POSITION OR POLICY.

SiTE1W ILDIUNIIENSIEGNS liN IIIiIG::B3IR IEDUDIca,'rum

A" LICATIONS OF THESCIENCE OF MEASUREMENT TO HIGHER EDUCATION

JuniusA. Davis

Everett H. Hopkins,Editor

U. S. DEPARTMENTOF HEALTH,EDUCATION, AND WELFARE

JOHN GARDNER,Secretary Office of Education HAROLD HOWE II,Commissioner ABOUT THE AUTHOR

Junius Ayers Davis is Director of theSoutheastern Office of Educational Testing Service, located atDurham, North Carolina. He received the A.B. degree from theUniversity of North Carolina, in mathematics; the A.M. degree fromTeachers College, , in guidance; and the Ph.D. degreefrom Columbia University in counseling psychology. As a member of departments of psychology,education, or sociology, he has held teaching appointments at anumber of insti- tutions, including Orange County CommunityCollege (N.Y.), Princeton University, Emory University, theUniversity of North Carolina, Rutgers University, BrooklynCollege, Univers:".y, and Columbia University.In 1957, he established, as Director of Testing and Guidance for the University System-Df Georgia, a cen- tralized research and advisory service for the seventeencolleges of that system. From 1958 to 1961, he was GraduateDean of the Uni- versity of North Carolina at Greensboro, andProfessor of Psychology and Education.In 1961, he joined the research staff of Educational Testing Service in Princeton, New Jersey, where,immediately prior to his present assignment, he was head ofthe Higher Education Research Group. He has served widely as a consultant onhigher education research, with current standing appointmentswith the Veterans Administration, Williams College, DukeUniversity, and the Regional Educational Laboratory for theCarolinas and Virginia. Dr. Davis' memberships in professionalorganizations include the American Psychological Association, theAmerican Personnel and Guidance Association, Sigma Xi, the Associationfor Institutional Research, and the American Educational ResearchAssociation. He has published widely in educational and psychologicaljournals, with recent emphasis on admissions research,criteria of educational development, and academic climate. TABLE OF CONTENTS

Page

About the Author iii

Foreword Highlights vi I.The Development of Selective Admissions Practices in American Colleges and Universities 1 A Brief History of Selection Practices in U. S. Colleges Admissions in the Fircrt Half of the Twentieth Century: Evolution of the College Board 20

The Prediction of Success in College 43

The Colleges and Admissions Today 57

II.Applications of Measurement in Higher Education for the Second Half of the Twentieth Century: Achievement and Prospect 68

The Emergence of a Science of Measurement 68 The Enlarged Concern with Student Input 86

The Enlarged Concern with the Educational Context 102 The Measurement of Impact of Colleges Upon Their Students 114

Directions for Future Research and Development 117

Footnotes 124

Annotated Bibliogra phy 133

Reactions

iv FOREWORD

(If and when this manuscriptis published for general distribution,the Editor will gladly prepare an appropriateForeword for the wider audience.) HIGHLIGHTS

Part I of this literaturereview provides historicalperspective for the development ofselection practices inAmerican colleges and universities, as well as areview of the development andstatus of measurement science in itsmost important routineworking applica- tion to higher education,namely, selective admissionsinto college. Part II begins with a moreintensive look at the developmentof general measurement science,and reviews its application,particularly in the last decade, toother uses in highereducation.

1.American colleges havebeen selective, covertly if not overtly, from the verybeginnings in the seventeenth century."Standardized" testing acrosscolleges for selection purposes beganonly in the current century, in response to bothsecondary school and collegeneeds. 2. Admissions testing in itspresent form began onlyin the late 1940's; the objectivetest for this purpose came into being for efficiency reasons,but it has perse- vered because of the ubiquitousrelationship between tested measures of scholasticaptitude and academic performance in college. A variety of attempts tosupplement measures of tested academic aptitude and measuresof high school per- formance with measures of othertraits (e.g., interests and motivation) have not improvedthe prediction of per- formance in college to the extentthat there the measures of these other traits are in commonuse. 4. Measurement as a sciencebegan only in the current century.It is marked with apreponderance of concern, during its first fifty years,with a focus on individual differences; only in the last decadehave measurement researchers in higher educationbegun to extend the science to the measurementof social and institutional forces. .vi 5. The lastten yearsmarks an explosionof interest in using measurementscience for studyof a variety of problems in highereducation.Significant factors in the emergenceof organizationswith this explosion are the avail- ' multi-college interestsand responsibilities, ability of substantialfunding for massiveefforts, and the use ofmission-oriented teamsof measurement research specialists. include the 6.Exciting newapplications of measurement broader study ofstudent input factors,of procedures environ- for measuringimportant forces inthe learning ments, and theprescription of elementsnecessaryfor a developinginsight into theimpact of colleges on students. for the decadeahead may wellbe a con- 7. The new look of cern withmeasurement as atool in the assessment interaction betweenthe individualand his learning environment, towardprescribing andvalidating effec- tive teachingprocedures for avariety of individuals, rather than as atool for only sortingout thosewho learn quickly andreadily in conventionalsituations and where successis measuredthrough standardgrading practices.

vii I. THE DEVELOPMENTOF SELECTIVEADMISSIONS PRACTICES IN AMERICANCOLLEGES ANDUNIVERSITIES'

A Brief Historof Selection Practices inU. S. Co llees2 In the beginning wasHarvard.

The model was EmmanuelCollege of Cambridge Uni%crsity. The year, 1638.The function of the newcollege was to insure aliterate and enlightened clergy ofnative sons after thoseeducated in England had wheezed for the last timethrough Old Hundredth; theinstitutional goal, as found in the StatutaCollegli Harvardini of 1642, wassimply:

" Considerato unusquisqueultimum finem vitae acstudiorum, cog- nitionem nimerum Dei et JesuChristi, quae est vita aeterna."The course of studies,standing virtually withoutchange throughout the first hundred years ofoperation, involved the learnedlanguages and their grammars, rhetoric,and theological andphilosophical disputa- tions. Latin was not only amajor subject of studies,but also the medium of communication, andstudents were forbidden to usetheir mother tongue within thelimits of the college.

In this context, and against someof the issues and complexities of selective admissionstoday, the first known statementof admis- sions requirements stands as areflection of and tribute to the 2 simplicity and austerity thatcharacterized New England in that period. The statutes of 1642, translatedfor the unenlightened by some unknown Puritan, read:"When any Scholar is able to read Tully orsuch like classical Latin ex tempore and make a ndspeake true Latin in verse and prose, suo (ut aiunt) Marte and declineperfectly the paradigms of nounes and verbes in ye Greeke tongue,then may bee bee admitted into ye College, nor shall any claime admissionbefore such qualifi- cations."

Thus, the criterion for admissions involved a standardof academic readiness, in terms of area and level of scholarly achievement.The standard provided guidance for those involved in preparingstudents for college; it also defined the elements of a simplesituational test that any college rector or tutor could administer acrossthe maple desk toward the identification, rather effectively, of prospectivestudents who could interact with, the academic world intowhich they were about to plunge.

Toward the conscience vs. intellect debate today it.is interesting to note that Increase Mather, president ofHarvard from 1685 to 1701, failed in an attempt to have a religious test inserted in the college charter; that event and ensuing bitter controversy resulted not in a revision of admissions criteria but instead the founding of Yale by the defeated orthodox Calvinists. Yet the first laws of Yale College 3 stated that " until they shouldprovide further, the Rectors and Tutors should make use of the ordersand institutions of Harvard College." Then, as now, the changingof educational practices was about as difficult to achieve as the movingof a graveyard; the underlying vested interests, though with somesmall differences in packaging, are in both instancesdeeply entrenched in the common dustof origin.

Therefore such blissfully defensibleadmissions criteria as those cited were maintained by and largethrough the eighteenth century by Harvard, and emulatedwith only minor variations by the twenty-two other institutions of highereducation that had appeared by 1800. Although these new institutions were aproduct of sectarian differences and regional concerns, their reliance onmother Harvard for faculty as well as a model of necessarystatutes, and on the grammar schools that existedsolely as college preparatory institutions, prevented drastic; revision from appearing necessary.Other subjects such as arithmetic and the sciences had not yetmade substantial entry into the curriculum of either college or grammarschool. The communality in origin, purpose, and curriculum(in which no flexi- bility could be either afforded or tolerated) prservedthe uniformity in the stated requirement (although some variation amonginstitutions in the required Latin authors had crept in by1800, and some iastitun tons outside the Boston area had added a requirementfor the rules

of vulgar arithmar:-.;).Still another factor which permitted the standard 1 4 requirement to stand without muchmodification was that there was " a general laxity of enforcementof the stipulated regulations for admission, and the[ora. examination was apparently a flexible and

informal affair."3

For America, the last half of theeighteenth century was a period of sweeping and dramatic change. The frontier,principally notable before for bears and buffalo, began to acquirebarrooms and bawdy houses, and the way was paved for the exploitationof the rich natural resources.Political independence came about in a climatewherein those most successful as architects for thebreakdown of class and caste, or for equality of access, won theresponsibility of managing the country. The state, with somethingequall: commendable but considerably more immediate to offer than the Church,experienced a separation from theChurch, and the Great Awakening was on.

But colleges then--as are colleges sometimesnow--seemed loathe to change. Their scholars and tutors went onchattering in Latin and Greek, and the curriculum remained virtuallyunchanged until other circumstances to be examined laterbroke through with the Civil War. The real impact of the new order thatwould ultimately play the major role in the development of admissions requirements was on new forms ofpre-college or non-college education. 5

First came the academies,probably starting with Philadelphia Academy in 1753. BeyondLatin and Greek theyoffered such subjects as English grammar,geography, algebra, geometry,natural philosophy, astronomy, music, composition,oratory, bookkeeping,logic, and virtue.The academies saw no reasonapparently why the minister- tutor or the grammarschool should be the solespringboard for entry into college; someof theirstudents not only desired to continuetheir education at " higher" levels,but also sometimes excelledthose who entered by other routes. Thecolleges responded by expandingand intensifying the specification of thetraditional subject masteries for admission, and by graduallynoting (and requiring)achievement in some of the newsubjects. By 1807, for example,Harvard' s require- ments read: No one shall be admitted, unlesshe be thoroughly acquainted with the Grammar of theGreek and Latin lan- guages, in the variousparts thereof, includingProsody-- can properly constmeand parse Greek and Latinauthors-- be well instructed in thefollowing rules of Arithmetic, namely, Notation, simple andcompound, Addition, Sub- traction, Multiplication, andDivision, together with Reduction and the single Rule ofThree; have well studied a Compendium ofGeography, can translate Englishinto Latin correctly--and have a goodmoral character. Each candidate shall be examined in theGrammar of the Greek and Latin languages, and in any partsof the following Greek and Latin Books, with everypart of which he must be acquainted, namely, Dalzel° sCollectanea Graeca Minora, The Greek Testament, Virgil,Sallust, and Cicero' s Select Orations.

The second impact of the new movement wasthe emergence in 6 the nineteenth century of thepublic high school. Beginning in1821 as the " people° scollege," these institutions werenot initially designed to provide college preparatorywork, but rather practical terminal courses to ensure aliterate population.But as tax-supported education and insistence on stateresponsibility in ensuring equal access to educationspread, college preparatory subjects wereadded. Bowles, 4 in his review of " the evolutionof admissions requirements, " attributes great significance to the Kalamazoo caseof 1874, wherein

" it was held ...that the state could act within its rights°to furnish a liberal education tothe youth of the state in schools broughtwithin the reach of all classes.' "The forty high schools in 1860 grew to 2,500 by 1890.

Broome° s review, though restricted tothe major colleges, gives the reaction of the institutions of highereducation over the period from 1800 to 1870. Not only were thereinvasions of new subjects in the preparatory experience, but also theintrusion of new teachers-- many of whom, in the rapidexpansion of lower education, did well to keep a jump ahead of theirstudents. No longer content that appli- cants " make and speake true Latin,"Harvard examined, in 1869, candidates for admission to the freshman class " inthe whole of Virgil; the whole of Caesar° s Commentaries; theOrations of Cicero, included in Folsom° s, Johnson° s or Stuart° s edition;Latin Grammar, including Prosody; and in writing Latin."Some hundred miles away, 7

Yale examined students the same yearin " Latin Grammar, including Prosody; Sallust--Jugurthine War, or fourbooks of Caesar; Cicero-- Seven Orations; Virgil--the Bucolics,Georgics, and first six books of the Aeneid; and Arnold° s Latin Prose Composition, to thePassive voice (first XII chapters) ."The requirements in Greek grew even more strenuous and varied.Yet the college curricula had changed little up to this time. The admissions procedures mostexactly reflected the harsh look downward into preparation rather than the hopeful look upward into promise; admissions requirements had become not so much a tool for the guidance of pre-college work as a weapon to impose the college perception of what indeed good preparationhad to be. The extensiveness of the new requirements may also have reflected a tactic for resisting, in the name of quality, the diverse curricular innovations outside the traditional college preparatory courses.

The War Between the States, the technological revolution, and the increasing clamor for public education, all had a profound effect on the 200-year-old pattern of college curricula.In the middle of the nineteenth century, the state of Massachusetts had withdrawn its financial support of Harvard when a committee of the state legis- lature found that an outdated curriculum failed to meet popular needs .5 Enrollment problems attributable to the war did not ease immediately following the war' s end. The prospective students 8 wanted something other thanthe theoretical and philosophical excur- sions backward into theclassical world; they wereanxious to learn how to exploit new ideasand techniques in the presentand emerging world. Though new colleges werebeing developed in greatnumbers by a variety of sects, it wasthe sect, not the mass ofapplicants, that desired an educatedclergy, and after a briefperiod most of these institutions collapsed becauseof financial problems and wantof students.

As early as 1830 there hadbeen some experimentationwith new college curricula.In that year Columbiainitiated the " Scientific and Literary Course" which,according to the statutory enactment, was establishedwith a " view of renderingthe benefits of education more generallyaccessible to the community."For admission to that three-year non-degree program,students were required to have a grammatical knowledge of French, aswell as meeting the usual --i'equirements in mathematics and geography.In the 1850' s Brown, Harvard, Yale, Dartmouth, Rochester,and Michigan introduced new non--classical degree programs. However, asthese came because of pressures from outside thecolleges rather than from pressures within, both curricula and admissionsrequirements (the latter in the pattern borrowed from theclassics and now synonymous withquality or academic orinstitutional respectability) developedwith an infinite variety of specific preparatoryprescription. This was the college' s way of attempting tomake such new programsreputable in its own eyes, or worthyof its attention and grudgingblessing, with a new degree (Ph.B. at Brown andYale, and B.S. at Harvard, Dartmouth, Rochester, and Michigan) as theface-saving rationalization to the higher academic community.

But both the new programs and theiradmissions requirements need to be viewed in terms of a majorprinciple stated byBowles:6 " The enforcement, or qualitative aspect,of entrance requirements is determined by higher education in responseto applicant supply and demand and with little or no reference to theattitudes and objec- tives of secondary education."The new college programs were not so much founded on deepconvictions as to the needs of a new society, but on urgent pressures to capture enoughstudents to permit the insti- tutions themselves to survive.This observation is of tremendous importance to those who would keep a sanehead and a perceptive

eye on the modernstruggles between preparatory and collegeforces; it is also fundamental to those whowould understand that the admis- sions problem may represent at first blushthe tool for repairing 'difficulties, but that it then becomes a dilemmawith one horn piercing the maintenance of collegiate or qualitativestandards and the other horn firmly planted in the necessity to adjustlevels to permit suffi- cient enrollment. 10

Thus, the explosionof publicedu6ation, the enrollmentand survival crises in thecolleges, and thetechnological revolution sounded the deathknell of the ColonialCollege, with the WarBetween the States and the ten yearsthereafter affording theperiod of break and transition. Oldinstitutions that wouldcontinue, and newcolleges to that would assume afirm foothold, wereforced to be responsive the non-classics,the mushrooming sciencesand technologies, sup- porting mathematics,the modern foreignlanguages, and the other subjects introduced bythe secondary schools.

The emergence of themodern American collegein 1870 carried with it the colonialadmissions patternstransmuted to the newsub- jects. With collegestudy today rapidlybecoming the prerogative of all, any modernscholar of more existentialthan historical bent who believes currentcontroversies and problems area+, a peak never before experienced needsonly to listen to thehue and cry of the period from 1870 to 1900 inAmerica. In his account(which reflects the biases of the privatepreparatory schools oracademies), Fuess has stated: For the preparatoryschools the uncertainty wasboth ludicrous and tragic. As[Nicholas Murra0 Butler said, " If Cicero wasprescribed, it meant in oneplace four orations and another six,and not always the samefour or the same six."When some collegesdemanded Greek Composition or LatinComposition, a school' sclassical department had to form specialsections to meet the need. Each college,furthermore, held its entrance examinations to suit itsconvenience, with theresult that the time scheduleof a school like St. Paul' s or Newton High School duringthe spring term was disrupted. One such group of examinationswas set on the dayof a school° s most important baseball game,and the local protests were violent.Dr. Cecil F. P. Bancroft, prin- cipal of ,Andover, complained patheti- cally in 1885 that "out of everyforty boys preparing for college next year we have morethan twenty Senior classes." ... The written[entrancg examinations themselves, often dictated hastilyby professors w...th small knowledge of studentpsychology, were unscien- tific and varied in difficulty from yearto year and from college tocollege.7

For the public schools, the situationin readying students for college must have been even moreludicrous. The high schools, showing the usual tendency of newpublic institutions to try to be all things to all people, did take onthe task of preparing students for the traditional programs of classicalstudies as well as for the newer, more practical areasof college work or for entrydirectly into work. The older, more traditional areasprobably attracted both the better students and the betterteachers; the newer, more prag- matic areas fared less well on both counts.These newer areas also sustained more suspicion from the collegesand entertained greater efforts, through admissions, to controltheir quality. The high schools grew inArnportance, however, andthis new market produced the observation of Broome at the start ofthe twentieth century that "the history of college admission requirementsfor a quarter of a century has been a series ofconcessions to the high schools." 12

These concessions werenot easilygranted; indeed, to givetoo the much emphasis tothem is to underrategrossly the impact of college, principallythrough its admissionsstandards, in providing the the secondaryschools with a mark atwhich to shoot. Even nature of the teststhemselves may havehad some desirableimpact, for as the rangeof facts to besampled on admissionsexaminations increased, the collegesnecessarily began tofocus on more general evidences of learning.Broome noted " theemphasis placed on sight translation in the languageexaminations, the growingimportance of

English composition,of the solution oforiginal problems in geometry, and of independentexperimental work in science."As a result of these tendencies incollege admissionexaminations, heconcluded: " There has been asignificant revolution inpreparatory schoolmethods of teaching, ashifting of the emphasisfrom stultifying memoriterwork thought and to that morequickening sort whichcalls for independent constructive ability.'"

There were, however,other solutions to theproblem of how to transmit the mold ofearlier testing in theclassical requirements into the broader arena ofthe expanding body ofsubject matter. One pro- cedure was the adoptionof a variety ofalternatives in the subjects

on whichstudents might beexamined. A bolder plan wasinitiated by the University ofMichigan in 1870. TheCalendar for that year stated: 13

Whenever the Facultyshall be satisfiedthat the pre- paratory course in anyschool is conductedby a sufficient number of competentinstructors, and hasbeen brought up fully to the foregoingrequirements, thediploma of such school, certifying thatthe holder hascompleted the pre- paratory courseand sustained theexamination in the same, shall entitle thecandidate to beadmitted to the university without further examination. One can well imaginethe furrowing ofshaggy brows inCambridge, the New Haven, andPrinceton, particularly asIndiana University, University of Wisconsin,and the University of'California 'followed suit in the nextfifteen years (the Easterncolleges had alreadybegun the "certificatesystem," whereby thoseparticular principalswhose wisdom and rigor werecertain from the factof graduation fromthe mother institutions weresometimes allowed tovouch for their candi- dates). The Michigan or"diploma system" wasrecognized as superior to the IvyLeague "certificatesystem" even byPresident Eliot of Harvard,who saw the attendantprocedures of the former-- involving inspection ofthe secondary schoolby the college faculty-- as a means togreater communication,interaction, and stimulation. Thus, the diploma systemassumed the previouslynoted function of admissions requirements asguidelines for preparatorywork, yet it

may haveachieved this more byfriendly cooperation thanby the super- imposed threat of the test orexamination standard.Whatever stand-- or strategy--anymodern critic would beinclined to take, Michigan, Indiana, Wisconsin, andCalifornia seem one hundred yearslater to be viable institutions ofhigher learning. 14

The last.quarter of thenineteenth century was to see animportant characteristic of Americaneducational systems develr)p--predominantly as a functionof the diversity of admissionsrequirements and solutions. 'This was the formal organizationof professional associations repre- senting both colleges andsecondary schools. The colleges,isolated from one another, harboringdelusions of self-sufficiency, orpriding themselves on their own particular brandsof wisdom, were not in any mood .to ease the admissions preparationbeyond relaxing standards when necessary to keep classes open(it is a curious fact that no instance is known where professorsclosed their books and went home for failure of available Students to meet apredetermined standard of excellence).This did not help the problems the secondaryschools were facing. Thefirst school-college organization wasthe New G. England Association of Colleges andPreparatory Schools, established in Boston in 1885 with the aimof " the advancement of the causeof liberal education by the promotion of interests commonto college and preparatory schools." What theseinterests ware became apparent from the outgrowth from this organizationin 1886 of the Commission of Colleges in New England on EntranceExaminations, representing all but five of the colleges in NewEngland, with its more precise aim " to devise means for securing greateruniformity in college admis- sions examinations."By 1897 there were twenty-threecolleges and other educational associations devoting timeand attention to the 15 problem of securing a workableuniformity among colleges.

These different organizationshad varying degrees of success within their regions of influence,but they met with enough success to demonstrate thatuniformity within regions was notsufficient and to attract national efforts.The National Education Association, an organization more representativeof public school interests than any other, appointed in 1892 the "Committee of Teri" to look at the problem on a national scale. Therecommendations of this group were studied by anappointed " Commit.,(?.e on College EntranceRequire- ments, " which involved nearly150 experts in both secondary and higher education working together for morethan four years. Their final report, in 1899, is a masterpiece ofeducational architecture. That is, it did not attempt merely toimpose a prescription for entrance requirements, but also attempted tostrengthen through guidelines for the secondary schools their preparatoryefforts (e.g., " we recommend an increase in the schoolday in secondary schools, to permit alarger amount of study in school underschool supervision"). The uniformity was not to be gainedby a common prescription, but through a system of units among various common subjects,with colleges to name the most crucial options withinthese units.

The work of the NEA and the regional groups is atestimony to the seriousness of the problem, and theirrecommendations, in terms of 16 sound. But what what was known atthat time, wereboth sincere and been fiery guardian ofsacred standards on anycollege campus has he is known to demur to acommittee representingthose very agents to higher dedicated through hisstandards to snareby the ear and lead things? It remainedfor the recommendationsto be implemented.

of At the very firstmeeting of the NewEngland Association Colleges and PreparatorySchools in 1895,President Eliot ofHarvard idea that had suggested thenotion of a commonexamining board, an had fallen on deaf earsamong hisHarvard faculty in1877.Professor Butler, later to becomePresident of ColumbiaCollege, introduced in that faculty in 1893such a resolution,which was passedby a unani- examine and copy mous vote.But other institutionsdid not rush to Butler to transport the Columbiarequirements, and itremained for Dr. the notion to ameeting of theAssociation of Collegesarid Preparatory Schools of the MiddleStates and Marylandin 1899. Thatassociation resolved to urae theearly establishmentof a joint collegeadmission examination board, to exactagreement amongthe member colleges as to eachsubject required by two or morecolleges, to holduniform examinations in June ofeach year, to empowerthe board to name secondary schoolrepresentatives to servewith it, and to request the member colleges toaccept thecertificales issued forsatisfactory performance on the testsin lien of theinstitutions° own examinations. Dr. Butler gave notonly of his counsel,but also space forthe offices 16 what was known at thattime, were both sincere and sound.BID: what fiery guardian of sacredstandards on any college campushas been known to demur to a committee representing those very agentshe is dedicated through his standardstosnare by the earand lead to higher things?It remained for the recomendations to be implemented.

At the very first meetingof the New England Associationof Colleges and Preparatory Schools in 1895, President Eliotof Harvard had suggested the notionof a common examiningboard, an idea that had fallen on deaf earsamong his Harvardfaculty in 1877.Professor Butler, later to become President ofColumbia College, introduced in that faculty in 189such a resolution, which waspassed by a unani- mous vote.Butother institutions did not rush toexamine and copy the Columbiarequirements, and it remained forDr. Butler to transport the notion to a meeting of theAssociation of Colleges andPreparatory Schools of the Middle States andMaryland in 1899. That association resolved to urge the early establishmentof a joint college admission

examination board, to exact agreement amongthe member colleges asto each subject required by two or morecolleges, to hold uniform examinations in June of each year, toempower the board to name secondary school representatives to servewith it, and to request the member colleges to accept thecertificates issued for satisfactory performance on the tests in lieuof the institutions° ownexaminations. Dr. Butler gave not only ofhis counsel, but also spacefor the offices 17

Exami- of the Board, and onNovember 17, 1900,the College Entrance nation Board wasborn.

In concluding hishistorical discussionthree years later(through argument that shouldbe examined in theorigival), Broome showed that recognition ofdangers resident inthe new system permitted a outside way aroundthem from the verybeginning. For example, an group taking over aresponsibility of the individualcollege had to do a creditable jobboth in constructingexaminations and inevaluating performance if the new system wereto survive.An examination writer in a college, inthe press of autumnbusiness, could use hastilydic- tated exams and browsebriefly through them, but nowhe must prove his mettle and academicintegrity. The secondaryschool representa- tive at the Board, sittingwith his college counterpart,could neither show weakness to thatcounterpart nor mercy tothe candidate. Uniform statement of thedefinition of admissionssubjects was assured, together with a meansof enforcing thesedefinitions. And finally, it was recognizedthat each collegecould preserve its own brand of integrity by doingwhat it wished, with thequantitative results: the Board would reportlevel of performance,define (of course) a " passing" level,but leave the collegefree to demand higher or accept lower levels.That these assumptionsproved viable was demonstrated in 1966 whenthe College Board, nowwith over 630 member colleges, expended morethan $18 million to providetesting 18 programs, a variety ofadmissions-related services, conferences and publications, and a varied program of researchand development (with almost $1 million invested in the lattercategory).8

It would be well at this point to summarizethe functions and roles that admissions requirements and procedures served intheir evolution to 1900, the problems they solved and theproblems they raised, and the forces controlling their establishment.Prescriptions for selective admissions started as a gentle set ofguidelines for preparatory agents toward the specification ofessential levels and the nature of ptior learning needed for orderly progression in the new learning. -environment. With the appearance not only of formal pre- college institutions to conduct preparation but also of a variety of these institutions, admissions requirements moved from providing guidelines to influencing, indeed controlling, the content, quantity, and quality of pre-college studies. That this responsibility could not effectively remain with the colleges has been seen as a product of upper academic conservatism and resistance to change, particularly that from the addition of new disciplines or content areas, of economic and social pressures on the colleges for survival, of the growth of knowledge and technological change, of the weight of the, public school population, and of the social and economic utility of the emerging subject matter involved in the people' s colleges or public high schools. The administrator of the blow for quality received a heavier one in return. Thepreparatory schools,particularly the public schools, emerged inthe driver° s seat, and the mosteffective resolution of the matter ofcollege-influenced qualitative standards came not fromthe rigid imposition ofcontrols through screening pro- cedures but from efforts ofcollege people working with prepar-atory people toward operationalcommunalities.

In the struggles to 1900, one mayalso discern the fact that admissions requirements werefrequently perceived as a means for the college to maintain status orits own brand of reputability.Yet one man' s statussymbol is another man' s poison,and not all those institutions or vested interests amongthe faculties who desired status actually earned it.The pure weight of maintaining, yearafter year, individually constructed examinationsled frequently to inferior or faulty samplings of subject matter content;and some of the areas through which status could be expressedappealed only to. the clois- tered proponents rather than tothe public on whom they depended for students. Also those with aparticular item to sell sometimes found the market had changed.

Finally, it appeared that for administrativeconvenience as well as for orderly controlby lower education of the preparatoryexperience, the responsibility for designatingand administering the academic con- tent of the admissions requirementsmust be passed on, if any were 20 to survive,to.forces outside the individualcolleges. These emerged in two basic forms:the recognition and useof the evaluation of the student by the secondaryschools, and the establishmentof a new institution to represent, for acollection of colleges and preparatory schools, their common needsand interests. This institution,appear- ing as the CollegeEntrance Examination Board,showed signs of succeeding by efforts (1) to representthe two partners,(2) to rely on the bestscholars from the ranks of the partnersto determine sub- stantive content, and(3) to refrain from a common prescriptionof what content and levels eachcollege could or could nottolerate.

Admissions in the First Half of theTwentieth Century: Evolution of the College Board, The development of our selectionpractices in the first halfof the twentieth century is probablybest traced through thedevelopment of the College Entrance ExaminationBoard, with attention ultimately being given to those institutions itserved and those it did not(and, in the latter case, whatthese institutions did).Probably no single organizational or consciously contrivedadministrative force has had

a more sweepingimpact on higher educationthan has this organiza- tion.

Several factors in the initiationof .the Board have already been alluded to, but should be recounted.First, the Board grew out of needs for enough consistency incollege admissions requirements

- 21 that the secondaryschools might have areasonable chance of pre- paring any able studentfor a range of institutionsof higher education. The chief architects werenot only the presidentsof two of the most reputable and visible institutionsof higher education, butalso educa- tional leaders of greatsubstance who were recognizedthen as they are today. Theirplan called for involvementof the most respectable academicians irom both higherand lower education; theoperational focus was t.) help define, ratherthan impose or enforce,academic standards, and to providecareful, sound, and fairevaluations of student performance. The Board wasto be a memberorganization, with policy, control, andactivities to be determinedby representa- tives of participating collegesand secondary schools. Andlast, but not least, it would assumethe operational burden of testing candidates at locations acrossthe country, and relieve thecolleges of construction, administration, andevaluation pressures.

The early history and developmentof the Board has been traced in an intimate, folksy accountby Claude Fuess.9 Although many errors of fact creptinto this stream of personalreminiscences, and there are many important omissions,the flavor of the Board' s early operation in donated space atColumbia University comes through vividly. Nine subject matter areas wereagreed upon (chemistry, English, French, German, Greek,history, Latin, mathematics, and physics), and forty tests coveringvarious courses within these areas 22 were preparedfor administration tostudents at the close of the year in which they took theparticular courses. Guidelinesand standards from the appropriateprofessional associations representingthe dis- cipline were consulted wheresuch were available. A committeeof examiners was appointed and a" Committee of Revision"(the nine examiners with the fiverepresentatives of the secondaryschools on the Board) was established toreview together the firstexaminations produced. In June, 1901, 973 candidates weretested in sixty-seven centers in the United Statesand two in Europe. The 7,889 papers written were forwarded to theBoard, where thirty-nine carefully selected readers sat around tables inthe Columbia University Library to evaluate them.Both the anticipation of this choreand the fact of physical meeting required formalconsideration of evaluative guide- lines and procedures; the taskof reading, the drudgery, the points of debate, and the humorous answerssometimes encountered gave to the Board whatw6s to be its distinctive personalityand flavor for the'next forty years. This was the in-groupof scholars, faced with a reasonabletask to give them focus, and withthe responsibility of defining the true substance ofintellectual development in their dis- ciplines.

Not all institutions rushed to jointhe new Board. Only Columbia, Barnard, and New York Universityabandoned their own examinations that first year.Eliot' s faculty at Harvard voted with nodissent that 2 3 certificates, and Yale it was " inexpedient"to rely onthe Board° s the Board' s trusted only its ownfaculty to evaluatethe papers. But evaluation of the papersWas, if anything,obviously too severe to were judged permit the criticismof laxity(40.7 percent of the papers quite in the failurecategory), and theexaminations themselves were than most obviously superior,with the time andthought given them, of those producedby the individualcolleges.

three In 1903 -theBoard° s constitution wasamended to include Central, and newschool-college associalions(New England, North Southern States) toparticipate with theMiddle States. Harvard Reserve; joined as a memberinstitution in 1904,together with Western in 1908, Williams and Smithjoined in 1907,Dartmouth and Wesleyan Joining meant sending Yale in 1909, andPrinceton and Amherstin 1910. not tanta- an officialrepresentative to theannual meeting, and was mount to acceptingthe examination program.It was not until1915 tests as. that Harvard, Yale,and Princeton agreedto use the Board substitutes for their own.Even then, theseinstitutions wantedthe June) and Board to draw uptheir own tests forSeptember (rather than time. found that they had tofurnish their ownreaders at that busy

Though Fuess° accountdoes not say so,it wouldseem a good guess thatother factors were atwork to make theexperiment success- ful. The earlysubject matter experts werekey scholars fromkey 24 institutions; students electing the examinations camefrom the pace- setting preparatory schools (whose teachers hadbeen involved), and colleges began to take a second look when the best representatives of their faculties returned to join in the local policy debates.By 1910 the examinations were taken by 3,731 students, with 1,626 from New England schools, or 1,9 68 who desired to attend New England colleges.

There were, of course, mistakes made. The first examinations were not of even quality, and some that did not prove out toowell gave particular pause to the readers. A policy of announcing names of top-scoring students backfired (who, indeed, could make a 100 in history, asked teachers at the schools with no such candidate). The professional associations that had been counted upon for substantive advice as well as status proved in general to have no members really interested in secondary school or freshman-level certification, and consequently were of little service. There also proved to be no effective way to resolve the question of absolute but equivalent levels of achievement from subject to subject against the varying standards of the representatives of the separate disciplines.(In 1914, for example, only 32 percent of all the candidates in American history received a passing mark; the Board reviewccommittee con- cluded that the reason was inadequate preparation, but had this writer been a secondary school history teacher at that time, the 25 next 3,000 words would be devoted to demolishing that explanation.) And, for the first twenty-five years, the Board opeL'ated on deficit financing.

As interest grew, however, the Board found itself confronted

with demands for examinations in additional areas:in fact, in 1902, , examinations were added in Spanish, botany, geography, and drawing. By 1916 there were clearly too many bits and pieces to be manageable, and the Board resolved the issue by introducing " The New Plan,"con- sistthg of four comprehensive examinations all to be taken in the senior year. These were the tests requested by and tailored for Harvard, and agreed upon by Princeton and Yale; they were to be accompanied by the report of the student' s high school average.

Two aspects of the New Plan are important.First is the change in form of the tests. The general or " comprehensive" nature eased the problem of specific content (in reducing the number of tests required to cover a subject area to one), making test-making and administration more manageable. Also, and more important, the move involved a recognition that tests going beyond memory of specific factual content into the understanding of basic relationships or the ability to perceive new relationships might be more defensible. These were perceived, in that day of Binet; as a step toward power or mental ability tests, although they were more closely akin to today' s achievement tests 26 than to scholastic aptitude as weknow it now.

The second important aspectof the New Plan was theprovision, through recognition of the highschool average, for taking into account the judgment of the secondaryschool teachers personallyacquainted with the student and hiswork. This, too, was a move awayfrom the absolute faith in only that standardwhich the college representative might prescribe.

Although the Board continued tooffer the old tests and services along with those of the New Planfor a number of years, it wasthe

New Plan thatforeshadowed the shape of things to come.The reasons may have been the pureweight and unmanageability ofthe old examina- tions; indeed, the Board hadfound in the burgeoning numbers ofdiffer- ent examinations andcandidates a reflection, rather than aneasing, of the multiplicities that hadled in large part to its form3tion.Another reason may havebeen the underlying tnith that anyteacher, at any level, must have a personal sayand concern for how his students are to be evaluatedfor the work under his direction; or, to sayit simply, that the teacher cannotafford for others to do all histhinking and planning for him(shades of the modern criticisms aboutteaching machines!). But the most likely reasonof all is probably given in

Chauncey'sl° comment about this period some years later: But as time passed, thecolleges discovered that de- tailed mastery of a large numberof individual subjects was 27 not as important as it had been thought to be and that the school record was at least as good an index of success in colleae as the examination record.[Emphasis added.] In other words, the colleges were discovering that what they wanted were students who could perform well academically, and that the report of the previous teacher was as good or better for this purpose than the test. The age of prediction of performance was about to begin.

Even with the New Plan, there was a long way to go before approximating modern experience. Chauncey also stated in his 1947 review: These [New Plan] comprehensives proved to be good predictors of college success if--and a significant if--the student attended a school which organized its work by the College Board' s published syllabus and which gave con- stant drilling in essay writing on the prescribed subjects. In other words, the goal of using tests to determine ability to handle new material rather than simply reflect old specific acquisitions had not yet been reached.

This matter caused few problems at first, for before 1920 the Board had concerned itself with the prestige colleges and the expensive private schools from which these colleges drew the bulk of their stu- dents. The 1920° s, however, were the years of the postwar boom, which meant a consequent deluge for higher education. The deluge involved a mass of applicants from public high schools. These schools 28 and their students scarcely suspected the existence of the Board, much less were they familiar with its syllabi. Also, although the private schools were malleable in their dependence for existence on .the benevolence of the colleges, the public schools were in no posi- donor disposition to look to HarVard as the only truly divine source of wisdom about curricula.

As noted earlier, colleges in America of whatever quality have not been prone to wither away because of lack of high-level students when there is a large market of potential applicants at lower levels. Neither has it been easy to lower standards, and the colleges that were to remain with the best of both possible worlds needed a way to identify and attract that segment of the large market that would be most likely to survive or do credit to the college.

In 1924 Professor Brigham of Princeton Universitywas appointed as the chairman of a Board commission charged with developing a new test- of scholastic aptitude.This was, of course, a time when those psychologists who had worked during with the new mass tests, the Army Alpha and Beta, were back on their campuses, flushed with enthusiasm and seeking new fields to conquer. Brigham didpro- duce a scholastic aptitude test, cast in " objective" format.In 1926, the Board agreed to experiment with it, and administered it without charge, sending the information to the colleges for guidanceor research 29

It was a new educational considerationthat provided the first entry for Brigham' s new test.Harvard, Yale, Columbia, and Princeton decided that it would be desirable to attractstudents from other parts of the country, and so established regionalscholarship programs. The scholarship applicants, coming mainlyfrom public high schools, did poorly on the regular examinations of theBoard, and there was the additional difficulty that the Board examinations,with results not available until July, came toolate.

Henry Chauncey, Who was responsible for the scholarship program at Harvard, was attracted both by the firstexperiment with the new objective tests in the Carnegie study in Pennsylvania by Learnedand Woodn and by Brigham' s work. The new objective tests could be administered quickly (the conventional tests of the Boardthen required a week ofwriting) and handled efficiently; they promised, in the absence of reliance on specific subject matter, a fairbase regardless of thk, secondary school program. Thus, in 1937,Harvard, together with Yale, Princeton, and Columbia, asked the Board to prepare a special series of objective examinations to be used inthe selection of scholarship students. This first scholarship seriesconsisted of a Scholastic Aptitude Test(with verbal and mathematical components) and a battery of achievement tests of whichthe applicant was to take 30 three.

In those days, however, award ofscholarship and award of admission were two different administrativeactions; the latter still depended on the regular entrance examinations.Brigham's work had tested, of course, the ability of his tests toindicate level of future academic performance; but the scholarshipofficers joined vigorously in this enterprise when some oftheir prize scholarship winners were turned down a few monthslater in the admissions office. The seemingly esoteric statisticalwork of Brigham was now studied carefully; someexperimentation was agreed on in the colleges to test the new type of index of promiseagainst the old type of index of readiness. And,against the criterion of level of academic performance in college, no advantage ofthe old-type tests could be found.

How long it might have taken for the Boardand colleges generally to take such evidence into account onits own weight is not known, for it was another train of events that led tothe program as we know it today, a train which started moving at 1:07 p.m.Eastern Standard Time

on December 7, 1941.The colleges responded immediately to the war pressures withround-the-year programs and new classes to be admitted in June, 1942.In the resulting clamor, the old-style essay testshad 31 to be abandoned(it was assumed for theduration) in favor of the more efficient model of Brigham.Second, the Board, withi psychometric laboratory now at Dr. Brigham'sPrinceton University, felt compelled to contribute all itstalents and facilities to the wareffort, and thus made the laboratory available for governmentneeds for tests. The annual reports of those yearsshowed, both in dollar income andin numbers of candidates tested for variouscollege or non-college war training programs, that this workboomed. But not only was capability and momentum acquired in thelaboratory:the technicians acquiring the heavy testing experience insisted onresearching the effectiveness of the tests, and bothimproved their product and acquired substantial proof of results.

The essay tests, except for recurring attemptsto contrive English Composition tests that could be gradedreliably (a quality the new young breed at thelaboratory insisted upon), were never to return. The patter of summer excursions ofreaders to Columbia had been broken, and the numbers now involved were tooheavy to permit the former style of operation. Not only hadthe test technician matched the old subject matter expert at his gameand against his ultimate performance standards, but he had alsocontrived a more efficient system.

Other advantages were noted. In theReport of the Executive 32

Secretary of the Board in 1945, JohnStalnaker observed: In 1900, the problem was to keep out allwho had not undergone very specific preparatory training. Such training was necessarily restricted tothose able to attend the few secondary institutions which devoted themselves to the whims of distinguished universities. Today the problemis to attract the intelligent, apt pupilregardless of where or how he got his training and almost irrespective of what his school has been offering on its curricularmenu.12 Following this vein, Stalnaker then noted that the Brigham brand of test was an "accurate index of pupil ability, rather than a meansof controlling the curriculum."Again, the bases for Bowles° 1956 observation that the secondary schools and candidate supply were the most powerful factors in admissions procedures can be detected. But the fact remained that the new tests gave the colleges what they wanted--good students. Chauncey discussed the new look in a1947 speech delivered before a midwestern group of college admissions counselors: But we have diverged not at all from our original goal of assisting adrhissions officers to do their job and of con- serving and enriching the human resources of this country by helping to ensure that the best students go to college.... The fixed star which guides our present course I might call the star of freedom. Or I should say freedoms, for there are three. Freedom from bias infavor of any group of stu- dents, freedom of subject matter and teaching methods in the schools, and for the colleges, freedom to use the scores on our tests as they seefit.13

To maintain an examination program of this sort required a different kind of staff from that employed by the Board before 1941; also demanded was a substantial ongoing research operation.The laboratory in Princeton 33 had the nucleus of staff andstyle (in 1946 the Princeton laboratory had more than one hundredemployees, while te Board's office at Columbia had become a virtual maildrop).It remained for a Harvard president, Dr. James B. Conant,to discover once morethe bold and creative solution and drive it home. Aschairman of the Carnegie

Foundation's Committee on Testing in 1946, heproposed that the testing functions of the Board, the American Council onEducation, and the Carnegie Foundation for theAdvancement of Teaching be con- solidated and placed with a completely neworganization, to be built from the resources of theBoard's Princeton laboratory.In 1947 the charter for the new organization, to beknown as Educational Testing Service, was granted by the State ofNew York; funds, equipment, the continuing non-college or government contracts,and staff were pro- vided or already held to assure a splendidstart.Thus-, Educational Testing Service, the " testing industry" ascritics then and now are likely to call it, came into being.For the Board, ETS would be used as its testing agent(to build, administer, and score tests, and to report test results) and wouldbe paid for services rendered.It would also conduct necessary research.

Clearly, the Board could now havefaded into the underbrush on Morningside Heights, leaving thebrave new breed of psychologists and statisticians to rule the entrancetesting business. But appoint- ments of great significancehappened. Dr. Frank Bowles became 34

Director of the Board,and Dr. WilliamC. Fels its secretary.Both Bowles and Fels were menof too great abilityand vision to let the passing of an era and afunction displace anyconsideration of new contributions. With the newETS free to handlethe mechanics as well as the theoreticalproblems of measurement,the Board could now turn its attention toadmissions problems andphilosophy beyond that of affecting the curriculumof the secondary school orfeeding only the vested interests ofprofessors in the variousdisciplines. These new areas ofservice are still very muchin process of formulation (some are dealt with in detaillater in this review), but afew examples here will suffice:the initiation of a programfor determining, by objective and confidential means,the financial capability ofparents, with the aim of providingguidelines for scholarship aids(1954); the initiation of a publications programfor disseminating information useful to college admissionsofficers and pre-collegecounselors; and the establishment of a grants programfor support of generalresearch (to be conducted by agencies orindividuals across thecountry), thus expanding skills or points of viewrepresented by ETS, its prime con-

tractor.In short, the Board turnedits attention to societaland mana- gerial problems of admissions,relegating the testing science tothe specialists.

The foregoing descriptionof the development of theCollege Board omits a large part of the pictureof selection for highereducation in 3 5 the first half of thetwentieth century. This isbecause the Board had traditionally concerneditself with a distinct classof institution, representing only arelatively small proportionof the institutions of higher .ducation (in 1950,less than 10 percent of themember colleges were taxsupported colleges oruniversities).True, the Board clientele were theinfluential, pace-setting institutions,important as models in every aspectof functioning; true also thatthese were the institutions concerned most directly withselection as a discrete administrative act (that could afford tobe selective), i.e., as afunction of the num- bers of applicants vs. the sourceand amount of operating fundsthat probably determine optimal size.But the great mass of tax-supported institutions, or those newer privateinstitutions that could not afford much selectivity during this period,need also to be examined:for these institutions, developing in manyinstances as a product of dis- tinctively American trends and needs,have accounted for the large bulk of our collegeg.raduates, or our pools of top-level manpower.

The public colleges and universities,and many late-coming private institutions hungry forstudents, have generally beendescribed as " open-door"institutions. .This is, of course, notentirely true, because admissions then and nowhave been controlled in at least

three ways:(1) by some operational conceptsthatdefine for whom the institutions are appropriate(e.g., for high school graduates,for sons and daughtersof taxpayers in a given state,etc.);(2) by other 3 6

more subtleforces that attract somekinds of students and dissuade others; and (3) by the amountof funds investedin higher education within a region (e.g.,supply and demand considerations).In short, these institutions were nottruly "open-door" any more than the concept of "educationfor all" would mean that college-levelstudies are appropriatefor all the population.Instead, these were the insti- tutions that generally relied onpubllc-defined standards attliezar)1. of admissions rather than onadmissions examinations defined and controlled by the college.

The most popular and universal "public" standard wascompletion of secondary school studies. As a furthersafeguard, some kind of accrediting standardswere also applied tothe secondary school, as a form ofguidance and control of that class ofinstitution (these were set, typically, bstate public-schooladministrative control agents or by a privateregional accreditingassociation). Since 1865 New York State had monitored and certifiedquality and level of secondary school preparation through their Regentsexamination system, thus focusing on the studentdirectly and the system indirectly.In most instances, the characteristicfeature of the accrediting standard

was its establishmentand control more by secondaryschool than by ollege interests.

How, then, were theindividual colleges free todetermine their

, 37 own goals andstandards? That such a system did notproduce a mass of homogeneous institutions, orinstitutions that varied directly as a function of the quality of secondaryschools in the drawing area, is evident from the great diversity of typesand levels of institutions that developed during the first halfof the century.

Some institutions were blessed with acombination of limited support, heavy pools of potential applicants,and an administration sensitive to societal needs as well as tointernal college demands from competent faculty.The best example of such an institution is probably the City University of New York, whosestudents, at the point of admission or of graduation, have stoodhigh on any criterion of competence or achievement that couldbe mustered. Other institu- tions have drawn on applicant poolsin an area that has provided a hierarchy of institutions for a varietyof needs or ability levels; the best example of this is probably theuniversity system in California, where ths university, the state colleges, andthe junior colleges have aimed at different subgroups ofhigh school graduates, defining these subgroups in terms of level of highschool performance (the grade average or rank in class).In most areas the rise of normal schools or teachers° colleges to prepareteachers during the public school boom in the early part of the presentcentury provided a second set of institutions that,because of the circumstances, had to attract and service adifferent level of student from those already 3 3 involved in the major stateuniversities or private institutions.But whatever the factors, by 1950(and indeed the same holds truetoday) there was probably no instancewhere within the boundaries of a reasonably heavy population areathe best student (in terms, say, of SAT score) at one collegewould not be in the bottom quarterof the student population of anothercollege equally accessible geo- graphically.In these cases, it may be thatthese institutions were not organized on a do-or-diebasis (per Harvard-likestandards) so that higher education becameavailable togvirtually everyone whc could complete secondary school.That such an educational system has paid off is evident from thesocietal roles the graduates of these institutions have played.

Still another charac+-xistic waythat qualitative standards were maintained is indicated clearly by theclassic Pennsylvania study of Learned andWood.14 In almost any institution of any size or complexity there has been a tendencyfor different kinds of students (in terms of level of ability) to be attracted todifferent majors or programs of study. Thisintra-institutional diversity, in large part a function of thesubstance of different fields of study(e.g., business management vs. theoreticalphysics), allowed many students to rise or gravitate to afield and level appropriate to the collegedepart- mental standards. 39

The most important andinfluential selection activityof these colleges is still to beidentified: that is, theselection that took place at the hands ofthe institution throughthe administration of its own requirementsfor satisfactory performance, orthrough the course-by-course performancerequirements maintained by thefaculty. Although the two qualities cannotbe compared, the range ofdiversity in level of attrition wasprobably as great as the rangeof diversity in levels of enteringstudents. At the midpoint ofthe century, some institutions failed to graduate,for academic reasons, lessthan 5 percent of those studentswho entered as freshmen;for others, academic attrition rates ran ashigh as 80 percent. To myknowledge, no one has made athorough study of the factorsassociated with these differential attrition rates;in general, those collegeswith a high quality of entering freshmenhave tended to fail smaller propor- tions, but there were someimportant exceptions. As in the caseof the Board-type colleges,applicant supply and demand,itself a partial function of institution supplyand demand, has probablybeen the most crucial factor.

What, then, did the first halfof the present century add to the theory and practice of selectionfor college? Clearly, the most important developments werethose associated with the concept of mental ability, and the perfectionof techniques for measuring it. Rather than viewing suitabilityfor college solely as a functionof 40 the content and quality of the student's preparation, those responsible for admissions practices began also to look at the student's "sheer power of mind."The focus shifted from the school and curriculum to the student himself. Reasoning tells us this could not have come about without some workable levels of quality being attained generally in the secondary schools; but other factors contributing to the use of the mental ability criterion were developments in the field of psycho- logical measurement, needs for greater measurement efficiencies as the numbers of students to be measured increased, the catalytic impact of World War II, and perhaps the realization that the difficul- ties in legislating quality and,content in the high schools by the college through selection standards were, after all, insurmountable.

This development shifted the burden, if not the responsibility, for defining admissions criteria to the test technician. To survive, the technician could not operate simply by his own whims and fancies; he had to take cognizance of the interests of those he served. Re- search was also part of his way of life, and much that had been taken for granted was subjected to painstaking analysis and scrutiny. Although then, as now, there were vocal critics among the academi- cians who feared gross inadequacies in the new tests, the technician found that in practice he was taken more seriously than he desired to be, and much of his time had to be devoted to preventing too much faith in his product by the consumers. 41 The advent of the measurement specialist, as we shall now call the technician, made selection a subject of scientific inquiry rather than a matter of academic debate. For the measurement specialist, the most appropriate model for this inquiry was regression analysis, or prediction.Thus, a corollary of the shift to the mental ability criterion was the formal examination of how well any admis- sions criterion, however defined, predicted later performance in college.This procedure provided a way to take into account the aggregate judgment of the members of a college faculty by relying on their later evaluation of student performance as the criterion for validating the new tests.

Other interests the measurement specialist brought with him were his emphasis on reliability of student evaluation procedures, and his proclivity for the "objective" format; subject matter achieve- ment tests still seemed reasonable for the task of predicting later performance; the colleges demanded them, and the measurement specialist found that he could make a contribution here as well through placing the examination questions in the "objective" format.

Another aspect of developments in the first half of the twentieth 0 century was the recognition, in practice, that the secondary school both maintained and merited a considerable say in the question of 42 selec- who should go tocollege. This, in turn,focused attention on of selections."15 tion for highereducation as aproduct of "a series in In a forthcoming paper,Dyer describesthe admissions process the following terms: It is a processthat has large consequencesfor the careers ofindividuals and for thecharacter of the society of which they are apart. Theselections may bedeliberate decisions by parents,students, and institutions,or they may be theresult of social,cultural, and economicforces outside the range ofindividual human choice.The manner in which anadmission system operatesis thus partly a product and partly adeterminant of theseforces.16

The role of theelementary and secondaryschool in the selection process wasfurther emphasizedby the great massof U.S. colleges coming to dependheavily on the simplefact of graduationfrom secondary school asevidence of eligibilityfor admission tocollege. hog-tied to a common, The colleges found,however, that they were not and homogeneous mediocrity,but that forces(some within the college through someoutside) other than thosecontrolled by the institution indi- its admissionsprocedures could be usedeffectively to permit vidual institutionalfreedom to definegoals and standards.Neverthe- less, the period showedthat student supplyand demand was apowerful force with which tocontend.

Somewhat at odds withthe foregoing, therebegan to emerge a status hierarchythat seemed toderive from selectivity atthe point of admission to collegeitself.It was the mostreputable and 43 distinguished colleges, those blessed withthe most esteem by the academicians as well as the general public andwith the most appli- cants, that employed the test systemsfor excluding applicants. .Selectivity beyond the minimal dependence oncompletion of secondary studies came to be regarded as virtually synonymouswith quality of institution.That quality of the entering student was related to the quality of the institution could also be rationalized bythe reasoning that the more able the student, the more rigorousand advanced college work he could sustain.Ergo, as supply factors permitted open-door colleges to become selective, they tended to adoptthe procedures of the selective colleges, not only because they weremanageable and available to imitate, but also because theseprocedures seemed to promise an attractive qualitative evolutionfor the institution.

Finally, the period surveyed saw not only the emergence of the measurement specialist, but also the turning of attentionof those more generally concerned withcollege admissions to the management of the specialist's products and to the broader societal problems this responsibility dictated. The College Board, representing the interests not of one institution but those of many institutions, could nowbe free to look more closely at national educational interests.

The Prediction of Success in Co lle e It would take several volumes to survey adequately, study by 44

published study or evenstudy-category bystudy-category, the college; and research concernedwith the predictionof success in published studies by aratio unpublished studiesprobably outnumber of about fifty to one.Before attempting toplace a few of these themes studies in somehistorical hierarchy orbefore moving to major speculate in the most promisingmodern work, itwould not be amiss to on the reasonsfor all this activity.

It has alreadybeen noted that asthe voice of themeasurement specialist was heard in ourland the time offormal research inselec- tion had come. Themeasurementspecialist could notstand, as did the academician, onhis definition ofthe essential.comppnentsof an area tobe tested; virtuallyeverything he did had tobe verified by placing the measureagainst an externallydefined criterion. Whether building a testof achievement, orof some kind ofability, behavior, or motivationalconstruct postulated tobe associated with demanded academic performance,the measurementspecialist's regimen formal validation andcross-validation studies.

A second reason,which drew its powerfrom the teaming up of the measurementspecfalist, the subject matterspecialist, and the admissionsofficer, was the needto determinehow well the measuresworked for eachinstitution, and toprescribe practice There has 'from the guidelinesprovided by thevalidation research. 45 been some variety in the student populations, predictive components, and curricula, thus frequently making these studiesof potential general interest. Also, as selection is practiced for a time,the range of the admitted portionof the population on the predictor varia- bles tends to reduce, making repeated studies appear important.

A third reason for this activity has been the difficulty in captur- ing more than 25 or 30 percent of the criterion variance in the nets of the predictors.It was the measurement specialist and his expe- rience that determined what a reputable level of relationship between ability or past performance should be; what he achieved appeared useful.But substantial unexplained variance remained; some indi- vidual students conspicuously performed much better or worse than the predictive indices indicated might be expected. The validity barrier, like the sound barrier, has been a limit which men fain would push beyond.

A fourth reason has been the infinite versatility of all of us, measurement specialist, admissions officer, teacher, clinician, or simply parent with child in school, to postulate a variety of traits or characteristics that would appear to be associated with academic performance.It would seem patently clear that factors other than native ability and prior achievement affect academic performance: study efficiency, motivation, intellectual curiosity, freedom from 46 extraneous concerns, special interests,traits of character, et al., ad infinitum.Sometimes it has been the psychologistwho believes he is zeroing in on some basic andfundamental moderator of behavior; sometimes it has been the laymanwho wishes to show the psychologist a thing or two. Thetriumph of certain insight can sustainthe crowning of empirical verification, and so thestudies have been conducted-- and sometimes reported.

A final reason, but not the least significant, maybe that valida- tion research in this area seems notdifficult or inconvenient to conduct. Applicants or students can becoaxed or coerced to submit themselves to an examiner, their later academicperformance may be copied from the records, and the investigator canthen look at mean performance ofsubgroups at different levels on his predictor measures. Given a brief excursion intothe jungle of elementary statistics, he may discover that he can become the grandesttiger by exercising himself through correlational analyses and testsof significance. To the resulting confrontation, the Little BlackSambos who edit professional journals have not infrequently yielded ablock of pages.

All of these elements are apparent in the first majorpublished reviews of the literature of prediction of academic success.Harris found, in 1931 and in 1940, a large number of studies and aconsiderable 47 variety of experimentalpredictors.1-7 Many predictors beyond only when intel- intelligence andachievement onesworked well, but For the times, ligence was ignored ornot controlledin the designs. horde this is understandable;Brigham, Terman,Toops, Otis, and a successful of other nownotables had translatedthe notion of Binet to Freudian outcomes; others nowwish to try the samefrom Watsonian or bases. But if theseother measures provedvalid, the bestexplanation predictor and was alwaysbecause of the mutualdependence of the by tests criterion on theunderlying cognitivefactor best measured were of mental ability.Harris concludedthat methodologically we the a pretty sorrylot, and arguedthat we should learnto profit from mistakes of others,if we could not contrivebetter designs of our

own.

The next major andcompetent publishedreview was that by

Fishman andPasanella,18 covering the periodfrom 1948 to 1957. Fishman°§ summary ofthat review in asubsequent paper speaks concisely of whatthey found: It would hardly seemto be too muchof an exaggeration to say thatnearly every investigatorof higher education has done a studypredicting collegeachievement or adjust- ment.It also seems that everyinvestigatgr has done only one suchstudy. What is the upshotof all this research oncollege selection and guidance?Unfortunately, it canall be sum- marized 'rather briefly.The most usualpredictors are high school grades and scores ona standardizedmeasure of scholastic aptitude. Theusual criterion isthe freshman 48 average. The average multiple correlationobtained when aiming the usual predictors at the usual criterion is approximately .55. The gain in the multiple correlation upon adding a personality test score to one orboth of the usual predictors, holding the criterion constant, is usually less than

The failures to improve much on prediction over scholastic aptitude and achievement measures can hardly be attributed to absence of ingenuity of the experimenters in seeking or contriving a variety of new potential predictors.First of all, it would seem that all psycho- logical tests appropriate for this age group, devised for any purpose, have been tried a few dozen (sometimes a few hundred) times as a predictor of performance in college. This applies particularly to the mass of personality inventories and the interest tests. Of course, the majority of these tests were contrived for other purposes.But further, scarcely any concept that would appear to be pertinent.to academic performance, efficiency, or satisfaction has not been transmuted into a measuring device of some sort. We have tried tests of study habits and attitudes, achievement motivation, reading efficiency, and the like.The physiological indices have not been ignored; here experimental predictors have ranged from rate of growth of testees to vitamin (vs. placebo, of course) supplements. A great variety of biographical or personal history items have also been explored, to.gether with factors drawn from the nature of previous academic experience (e.g., public vs. private school background, or special 49 educational treatments). Wnen oneexcludes those studies that do not provide controlsfor intelligence and past achievement,the occa- sional successes are almost alwaysmatched, on replication attempts by the original investigator orby another, with failures to confirm the previous findings.

The serious student of prediction ofacademic success should not accept this report offailure and seek some more hopeful profession without examining this literature in greaterdetail than is possible here. More optimistic modern reviews arethose provided byStein20 or Lavin.21These and the earlier literature doprovide some small pockets of hope.

One of these hopeful signs may be drawnfrom the occasional consistencies in the findings that are positive,if of small degree. Both the studies of biographical factors andof social and demographic variables point to some pragmatic underpinnings ofacademic success or persistence: attitudestoward school, having a vocational goal (males only), parental educational level (or kJerhaps socioeconomic status). A consistent superiority of public school students over private or military school graduates of equal abilityhas been found (e.g., Koos, in 1931;22 Davis and Frederiksen, in1955;23 Shuey, in 195624); the most plausible explanation of thisphenomenon is probably that advanced by McArthur in 1954 and1960,25 who suggests 50

that in American societyachievement motivationand upward mobility through educationalattainment is essentially amiddle-class and not an upper-classvalue orientation.If any consistent thread runs through the studies utilizingpersonality tests or inventories,it is academic orientation orachievement motivation.

Yet the relationships thathave been discovered aresmall, and for each consistencyseveral new challenges presentthemselves. One of the most soberingof these is the sortraised by a number of / investigators (e.g., Holland,1959,1960;26 Getzels, 1960;27 Davis, 196528) who, in examining the criterion ofperformance provided by instructors° grading, have foundevidence that the cooperative student4, the one who conforms tothe value systems of the teacher, or the convergent(as opposed to the divergent) thinkeris the one who may have thedistinct advantage whenperformance is evaluated. Getzels, who sees the conventionaltests of ability andachievement as well as theconventional criteria stackedtoward "selecting the college studentmech.anically as Nlaral sox_n_7er,"states the problem: "It is the convergent individualwho is the most ready sourceof manpower, the divergentindividual the best hope ofMan."29 (He concludes, meekly, that we needboth.)

Another sobering challenge, beyondthe notion of inadequacy of the criterion, is that provided by thequestion of whether we are

v,r1 1;7 51 trying to force onselection some of the burden weshould be shouldering in our teachingand academic goals. Were there agood

measure ofachievement motivation, should weselect students for college on that basis, orshould we use it to determine whatinstruc- tional activities or contrivededucational experiences promoteit? One may argue, as have ColemanandCureton30 who found statistical justification for the notion, thatintelligence is associated with achievement because, in effect, theteacher's tests or other grading practices provide mostly ahome-made (and perhaps somewhat specialized) second test of intelligence.If taken to extremes, this can be a terriblydamning indictment of testing forselection to college and for what goes on in college.Are we insuring, by selection, that whatever we do or fail to do bythe curriculum and our teaching, our students will emerge only with anunderlying innate or.long-before acquired facility to memorize, tohandle abstractions, to manipulate symbols? In grading, do we certify theiroriginal promise instead of our impact by repeatingthe predictor measure?

We could probably feel morecomfortable in these arguments if we had devoted the sameattention to the criterion that we have given the predictors.In one sense, the earlier concernwith predictors would seem justified because it would seembetter that the measure- ment specialist rely on competentoutside authority to determine what is good or useful. But then there arenagging studies such as those 52 byPage31on essay gradingby computer; he hasfound that a computer, primed to detect commonmisspellings, a: to count unusual words, number of dashes, or numberof words, can predict thecomposite evaluation by trained judgesof general writingability better than any single judge can predictthe consensus of hiscolleagues.Similarly, the studies of Klein andSkager,32through focusing on the criterion of expert evaluation ofesthetic products (artsketches), have revealed some simple,mechanical guidelines that canbe taught to secretaries in a " five-minute artappreciation course" and whichenable them to match the evaluation of theprofessionals.In other words, the atten- tion of the measurementspecialist to the criterion begins to seem justified: not merely to prove,after all, that his predictor was reason- able, but to help the expertrealize the limitations and sourcesof bias in his own evaluation as aninitial step in contriving better criteria. At any rate, the measurementspecialists are beginning to show that they have a contribution tomake on the other axis of the regression plot, and are gatheringthe courage, or the security, or the capability required totackle it.

There are some other generallessons in the literature of research on prediction of successin college.I have noted in other reviews several instances where theemployment of the concept of moderator variable has permitted reasonableimprovement ofprediction.33 The classic work here was done bySaunders34 and Frederiksen and 53 Melville,35 who found that appropriate interest scales of the Strong

Vocational Interest Blank predictedacademic success for non- compulsive students but not for compulsivestudents; before treating the data in terms of compulsiveness(the moderator variable), no relationship was apparent.Ghiselli36 has also devoted attention to this technique. The possibilities here arestaggering, for one can, startina with sex, postulate moderator ormoderator-of-moderator variables by the hundreds.

Another lesson has been that if trying one new predictor after another is not likely to work, or jf diverting a newclinical instrument to the purpose of academic prediction is afrustrating exercise, then long-term persistent efforts by a competent team may be the answer.

The work of French and hisassociates37on differential prediction of grades in college involved painstaking item construction for abattery of cognitive variables, careful identification of separate sourcesof variance through factor analysis, and, incidentally, aninterest test constructed specifically for the academic prediction task.The place of the cognitive factors is not yet assured, but French didfind the interest test adding to the prediction of grades and predicting satis- faction with courses as well. The work of Schlesser and,Finger38 with the Personal Values Inventory is another case in point where involve- ment over time with a tailored rather than adapted instrument seems to be paying off (although they have persisted infailing to control for 54 past achievement insecondary schobi).

There is also evidencethat raw empiricism, orthe blind search for expedients thatwork, is not enough;Messick, our country's young dean of personalityresearch toward academicprediction purposes, has stated the case in asignificant paper that shouldbe read in the original.39 His arguments and work attestthe need for careful pre- liminary theorizing or modelbuilding, and for exhaustiveconstruct rather than predictivevalidity research.

Another argument too well thought outto be ignored is thatof Fishman,40 who argues for a social-psychologicalrather than a simple regression model. Hemakes a nice case, for example,for looking at the similarities andthe differences in the tasksand valLes systems or environmentalcharacteristics of the secondaryschool vs. the college,in understanding the successesand the failures in predicting success in onesetting from success in theother.

Some exciting possibilities fromthe current work will be saved for the second part of this paper,where, after examiningbriefly other applications of measurementscience, I shall attempt to specu- late on the most promising leadsfor future study over the.varietyof potential applications. This is appropriate,if we accept as our criterion the fact that current selectionpractice has not yet added universally accepted new measures orpractices. We have seen the 55 reasons for this failure as stemming from too opportunistic a flitting from one new measure to another, too much blind empiricism, too little attention to (or too ready acceptance of) the criterion to be pre-

. dicted, failure to take the complexities of behavior into account through the moderator variable approach, too much reliance on the regression model to the exclusion of new models, and failure to accept the necessity of long, massive, painstaking work if success is to be achieved.It is accurate to say that no one, whether some measure- ment genius working in a university office or laboratory as did Brigham, or the several hundred specialists at ETS with its expenditures for research and development of more than $2 million last year, has any promising new mousetrap ready for our market or for import.

Before leaving this area, it is imperative to crystallize one more implication of the body of research:that is the need, stated most effectively by Messick,41 to consider the ethical bases of selection. If we find that Jewish applicants outperform non-Jewish applicants, shall we be preferential?If young people from broken homes appear less likely to persist in college, shall we discriminate against them? If our current tests and criteria are biased against culturally different subgroups of the population, are we justified in continuing them? Many of the personality variables tried are even more problematical, for they imply not only unreasoned bias for particular qualities but also abdication of the institution's responsibility for changing 56

behavior. Applicantsof considerableclosed-mindedness may seem less desirable and bedeucedly difficult to teach;the John Birch Society

or the Ku KluxKlan may accept themif Siwash U. considersthem un- desirable.But in another perspectivethese applicants couldbe con- sidered more as evidenceof failures of theeducational system gone before than as undesirablesfor further training. Manyqualities of personality are, of course,the product of extra-schoolforces; and there is not widespreadsympathy now in America forthe assuming, by the college, of atherapeutic role (to the point that we areexpe- riencing a closing or phasingout of the many universitycounseling centers that sprang upduring the war years). Thewriter stands strongly with those of theopinion that the function ofthe college is to teach, not to treat;but in many areas ofpersonality development, good teaching in thetraditional disciplines may be aseffective as the psychiatrist's couch,barring serious psychosis, and manyof our mostcherished academic goals haveliberalizing or personal style cbmponents.It would seem we are ondangerous grounds if we rush tooquickly to establish sets of good vs.bad personality traits for use in selection.The final resolution mayhave to involve not only the measurementspecialist, the subject-matterspecialist, a. and the admissions officer,but also the teachingtechnologist, the humanist (particularly thephilosopher), and the significant leaders in the architecture ofeducational systems.

::::*4101111111 _

57

The Colle es and AdmissionsToday, If we preserve the continuityin the historical tracing of the development of American highereducation, its admissions practices and its testing movement, we mustturn now to the period fromabout 1950 to the present. There are manysocial, cultural, and educational factors that have influenced highereducation and selection in the last two decades: the informationexplosion, the rapid development of educational technologies, theubiquity of the computer, the civil- rights revolution, the spread ofurbanization, the legitimizing of the college as a focus for research by socialscientists.But none of these forces has had the profound anddirect bearing on the institutions of higher education that increasedenrollment pressures have caused.

The enrollment pressures are, of course, afunction of the higher birth rates during (and continuingafter) World War II, as well as of the larger proportions of the populationseeking higher education. These larger proportions are, in turn,derived from the continuing expansion of belief in education, anational prosperity that permits later entry into work as well asbringing higher education into finan- cial reach of more people, and the factthat in America old institutions expand and new ones spring up if there is ademand from prospective applicants.

The best summary of the extent of this increasein the United 58

of States is probablythat given byDyer in hisforthcoming analysis Bowles43 college and universityadmissions.42 First, drawing from "the severity and U.S. Bureauof Census datafor 1960, he compares with its severity inthe of educationalselection in theworld at large American system:"

Table 1*

World and U. S.Enrollments asPercentages of AgeGroups

World U.S. World U.S. Educational level % % % % 1950 1950 1959 19 60, 95 Primary 37 93 50 85 Secondary 18 75 27 24 Higher 3 15 5

*Table taken from DyerMS.

Dyer also notes,from various data: Of the totalnumber of Americanchildren who entered the fifth grade in1956, 36 percententered college in1964 by normal progressionthrough the system....Of the cohort that startedthe fifth grade in1942, only 21 percent entered college whentheir time came in1950.... In 1964, 54 percent of thehigh school graduateswent directly on to some formof college work ascompared with only 41 earlier group. This rateprobably reflects, percent of the of to some extentat least, boththe greater pulling power higher education andgreater accessibilityof its institu- tions.... Since1950 the collegeadmissions curve has become steadily steeper: 517,000 entered in1950, 690,000 entered in1955, 929,800 entered in1960, 1,453,000 entered in1965, 59

and conservative projections say thatthe number of new freshmen in 1975 will be something over3,000,000.44

Another factor making this possible, evenreasonable, is the already noted diversity that has emerged inAmerican higher educa- tion.This has been documented by a number of people invarious ways.45It has been frequently noted in our country in the pastfifteen years that there is a collegefor every high school graduate, whatever his ability level.

Yet the demand for admission has not been evenlydistributed over the range of institutions. Some are morepopularly aspired to than others; some tend to draw from certain segments of the population in terms of ability level or socioeconomic status; some try toexpand (generally the public institutions), while others try to maintain their size (generally the elite private institutions) but exercise greater selectivity. Although some colleges have turned away as many as nine out of every ten applicants, others have needed additional stu- dents to fill classrooms and dormitories.

Perhaps, as I once noted,46 selectivity would seem a rather powerful way to manipulate the character of the institution in desirable directions. Perhaps selectivity is a distinguishing characteristic of the distinguished colleges and universities. Eble'svolume47 leaves little doubt that a status hierarchy related to selectivity exists among 60

U.S. institutions and goesbeyond athletic rivalries.This hierarchy does not exist solelyin the minds of thescholars and academicians, but is shared ingeneral by the Americanpublic, with the result that the tightening ofselection proceduresin a college sets off a geo- metrical progression of greaterand greater selectivity.With enroll- ment pressuresincreasing, the mostselective colleges havebecome of necessity even moreselective. A great manyof the established public universities, and manyother private colleges withrelatively fixed endowment, havesuddenly had the luxury of anabundance of applicants and have enjoyed someadditional freedom not to try to squeeze all aspirantsinside by the developmentof newer institutions-- notably the junior or communitycolleges, now being created at arate greater than one a week onthe average--or by thewillingness of many of thelower institutions on the statustotem pole to try first for greater capability throughexpansion.

For the institutions thathave long been selective,putting preva- lent and uniform focus forselectivity on one standard cognitivecri- terion, there has been arunning out of range on this measure.When a university enjoys morevaledictorians and more high scorers onthe SAT among its applicantsthan it can accept, howthen is it to sort? This problem has had twodifferent kinds of impact.The first is to make the search for additionalbases for selection seem more urgent; the second has been a return to morecomprehensive achievement Olownwo.

61

President testing. Onthe latter matter,the currentExecutive Vice of the CollegeBoard hasrecently commented: has While the numberof candidatestaking the SAT since 1959,Achievement Test grown twoand a half times decade of volume has betterthan tripled.Whereas the be termed "theaptitude era,"the first the 1950's might "the return to five years ofthe 1960's couldbe called The most dramaticdemonstration ofthis achievement." where trend has been inthe AdvancedPlacement Program, A the numbershave increasedfivefold in five years. throwback in a senseto thesyllabus-basedcomprehensive Advanced PlacementExaminations tests of the1930s, the school have served toreimpose restraints onthe secondary through the coursedescriptions forthem curriculum directly place- and indirectlythrough the prerequisiteto advanced mentwork.48

the For themid-range,traditionally"open-door" institutions, possibility thatthey enrollment pressureshave brought intobeing the selectivity too.It was may both growand begin toexperience some where the College this market towhichparticularly inthe Midwest College Board seemed morelike a New Englandproperty--the American that a husky Testing Programseemed to appeal, atleast to the extent being in the second nationaladmissions testing programgrew into of the University latter half of the1950 decade. In1957 the Regents applicants System of Georgiaenacted the boldrequirethent that all in that state for any of the seventeenpublic collegesand universities College would submit scores onthe ScholasticAptitude Test of the Board; these scoreswould be usedinitially in aRegent-sponsored colleges program ofresearch, but thehope was thatthe stronger 62 academically could be morerestrictive and moreefficient (through others couldaccOmmo- taking fewerhigher-riskapplicants), while the date those whowould profit fromless rigorous studies ordifferent programs, or serve asa provingground for those for whomit seemed unsafe to start in thebigger league.It is the movementof the mid- range institutioninto the prospectof selection, ratherthan increased applicants for institutionswhich have long beenselective, that has accounted for the boom inadmissions testing(from a volume of admis- sions program testsof 746,522 in 1959-1960to a volume of2,076,470 in 1964-1965,according to the annual reportsfor those years of Educational Testing Service;for the younger AmericanCollege Testing Program, the comparablevolume figure for the latter yearwas705,063).

For the new collegesand those that wouldremain relatively open, the selection testingmovement has takenanother relevant turn.It is precisely theseinstitutions, with theirwider range of applicants, or with theirgreater preponderanceof students of loweracademic potential in the traditional sense,for which the tests would seem most usefullydiscriminating among students.These institutions are alsofrequently infested with avariety of programs. Here,too, testing with the examination3developed for selectionhas begun to build up, with the emphasis onpre-college counseling orguidance, determination of specificneeds for remedial work uponentrance, or placement.In this spirit theCollege Board has producedthe 63

Manual of Freshman ClassProfiles for Indiana Colleges(CEEB, 1965). It is significant too thatthe Board is currently mounting amajor effort to develop a usefulbattery of tests for communitycolleges.

To look at present timessolely in terms of sele-ction test prac- tices is to miss one importantflavor in admission to college in America. This flavor comes from thecolleges and their admissions officers. Although these are the consumersof selection tests, it would be a gross error to imply that test- orachievement-based criteria are their sole concern. From areview of college catalogues and statements of admissions officers,I have summarized their con- cerns in the following way: Our question now is: what qualities doadmissions officers seek? Several classes of criteria canbe found from a review of such statements, orfrom studying the admissions process in a number of institutions. The most frequently cited class of criterion isthat which pertains to qualities directlyapplicable to academic achievement. Ability tests and past achievement arethe main ingredients, but some colleges extendwith these names (at least) of unusualtalents or such characteristics as "thirst forknowledge."In this class of criteria are all those personal attributes that academic manhas postu- lated in himself and his prized students. A second class of criteria has to do withtraits or quali- ties generally valued in our culture.These have little or nothing to do with the business of learning,but reflect the opinion of important constituencies of thecollege as to what constitutes glowing young manhood orwomanhood. Typical are qualities of "Christian commitment"in church- related colleges, "leadership" in colleges envious tohave a place in the sun throughtheir graduates, or "personal stability" in colleges without facilities or courage todeal 64

with troubled students. A third class of criteria has todo with practical advan- tage or necessity, orspecific needs for maintenance of the institution and its particular programs.Colleges with affluent alumni admit preferences foralumni sons; athletic programs are maintained; statequotas are fillecL first at public colleges. These can work in reverse: onepres- tigious private college has found apreponderance of its applicants Jewish, and has found it must restricthere more severely or lose itstraditional secular character. A fourth class of criteria has to do withthe hope of obtaining a balance of students fromidentifiable and hopefully meaningful subgroups. Here, thesearch is for variety that is meaningful in itself.The provincial college anxious to become more catholic may givepriority to dis- tant applicants; a college with manyurban applicants may seek, as Harvard has stated it does(Glimp and Wnitla, 1964), applicants from a rural or small town background. Some colleges have been bold(or confused) enough to call for a still more sophisticated varietyof students-- this time, in terms of a catalytic mix ofpersonal qualities. Extroverts may be balanced against introverts,four-letter men against shy, young,thick-lensed scholars of four- teenth century French poetry, and so on. Thestudent body is certainly a potent source ofstimulation; whether it can be manipulated effectively this way remains tobe seen. Yet, some students were admitted last year inselective colleges primarily because they added some sort of season- ing to the freshmanbroth.49

These criteria indicate, first of all,that American colleges believe fiercely that they have the right toselect their own students. But, of the classes of criteria., measurementspecialists or other educational research personnel have researchedonly the 'first, that pertaining to qualities related to academicachievement. Research in the other areas may be difficult, but their contentindicates a 41. 65 need for widening the scope of ,ourstudies.

Our final area for review of current thinking anar)racticc the United States dravtis from the burden of research,the trends in higher education, the concerns of admissionsofficers, anCi the new areas of involvement of the College Entrance ExaminationBoard.This has to do with the examination of selectionbroadly in terms of its societal implications, rather than what it seems to promise testby test or institution by institution.

There is no longer much vocal concern about legislatingquality and content of the secondary school through selection standards. Selection may seem a simple way to enhance the quality of an insti- tution of higher education (this has probably been so for manyyears), but that is now seldom disguised as an attempt to impose college standards on the preparatory agents.

In the broader sense, selection is more frequently viewed as a system (affected by many forces appearing from the time of conception to the time of entry into work) for exercising our biasesfor the control of societi. With higher education becoming more pervasive for our population, the question has become less that of "Who shall be edu- cated?" and more that of "How and where shall various individuals be educated?"This view has major manpower implications, and means that the study of selection should have one foot firmly on the channels _ ..""`7"

66 of access through which people pass,and the other on our manpower needs for people with variouskinds of training.

As our attention has begun toshift from the sorting that takes place at the point of admissions to thesorting that takes place at other points and by other forces, we havelearned that self-selec- tionthe s2lection of the college by theapplicantmay have been more influential all along.A more reasonable control thanconscious, deliberate sorting among applicantsfor admissions may be that which would augment pre-college guidance at thepragmatic end, and explore coHege image or other factors that attractand distract different kinds of students at the developmental end.These are goals of the Board's current guidance program; and theBoard has commissioned a major longitudinal study of channels of access(a report of pilot work has been published byTrent)50 that is now under way at the Center for Research and Development in HigherEducation at the University of California at Berkeley.

Last, but not least, is the emerging roleof the measurement specialist.It would seem that he can no longerafford to content himself with the construction of predictorsof human behavior, but he must also enter the business ofhelping define the criteria by which that behavior and his predictors areevaluated, and the char- acteristics of the institutions, teachers,and personal experiences in which new behavior patterns are acquired. In many cases,he is 67 finding that his knowledgeof his predictors, and the constructs on which they are based, areuseful in diagnosing the componentsof the criterion and the situationsthat produce people who can meetit. Although his knowledge of his ownlimitations has gsown, he has also in that process acquired somecogency among his owncult, and now, if he is good, hasdesigns on attempting to facilitatethe impact of the educational system onthe individual and on the society of which that individual is a part. II. APPLICATIONS OFMEASUREMENT IN HIGHER EDUCATION FOR THE SECOND HALF OF THETWENTIETH CENTURY: ACHIEVEMENT AND PROSPECT

The Em.ence of a "Science ofMeasurement"

The first half of the twentieth centuy. The turn of the present century saw notonly the formation of the College Entrance ExaminationBoard, a multi-institution agentthat could become a major consumerand purveyor of the output of measure- ment science, but also ageneral turning of attention to two new areas of inquiry that were to form the crucialbasis for a science of measure- ment. One of these areas ofinquiry grew out of concerns within the discipline of psychology that individualdifferences must be considered, that a search for modal or generallaws of behavior was not enough. Thus, the search was on for meaningfulhuman traits or attributes in which individuals varied, for ways inwhich these differences might be quantified, and for practicalimplications of particular human vari- ability. The tenor of the times just before1900 can be captured by visualizing Professor J. McKeen Cattell atColumbia University asking his class to stand with armsoutstretched as long as they could, scoring this little test with the clickof a stopwatch as weary muscles gave way and armsdropped, and then seeking arelationship between or'

69 scores andacademic performance(the results of this pioneer attempt to assessgeneral motivation werenegative).

Shortly after 1900 theFrench physician AlfredBinet, who had been concerned wiLhthe diagnosis of mentaldefectives and with how " bright" and " dull"children differed, found anopportunity in the Paris schools to attempt, throughtesting, a sorting ofchildren into regular or simplified programs.This first real test ofintelligence or general mental ability proved bothsuccessful and useful, andfound quick translation (in 1910-1916) andfurther development in theUnited States principally through the effortsof Professor Lewis Terman atStanford University, whc reflectedthe growing interest amongAmerican psycholo- gists in the generalpsychological development ofthe individual.

The second major areaof inquiry needed forthe development of a science of measurementwas that ofstatistical procedures for dealing precisely with arraysof measures.Sir Francis Galton, his interest aroused byDarwin's new theory ofdifferences among species, was not onlyconcerned with the inventionof ways to measure physical characteristics, keenness ofthe senses, and mentalimagery, but also mathematical waysof expressing differences amongindividuals-- of placing the personwithin a group, and ofdescribing concisely the group as a standardfor comparison. Galtonand other Britishers such as W. S. Gosset(a statistician for the GuinessBrewery) carried such concerns on tovarious extensionsof the two grand conceptsin 70 statistics:that of procedures for assessing the degree of relationship between concomitant measures or conditions, and that ofprobability and significance, where (in simplest terms) one is concerned with whether observed differences in measures are the result of error in measurement or of some more basic underlying condition.

Although in the other mainstream there have been occasional geniuses who have made a significant if lonely mark, trait definition and development of test or measurement devices has been, oddly enough, a product of hard times in a generally affluent and literate society. American psychologists have clearly dominated the fieldin both the invention and.the application of measures(Great Britain is just now, in the late 1960's, getting around to determining,through its Vice-Chancellors.' Committee on Entrance Procedures, if the objective American admissions tests have any relevance forthem). Also, history shows that the test development milestones in signifi- cant and successful production of human measuringdevices have come in periods when candidates were either widely available or there was a need to select them quickly(both of which place a premium on choosing without waiting for a test of performance overtime). Thus, the first boom in the construction of measuring devices came with World War I, when groups principally at the Carnegie Institute of Technology concerned themselves with the invention of measures for classification of servicemen toward assignment to training opportunities. 71

Psychologists involved inthat effort, such as EdwardK. Strong, Jr., Herbert Toops, or ArthurOtis, were to hold theirplace in the psycho- metric sun for the nextseveral decades.

We have already cited theadaptation of these efforts tothe college admissions problem byCarl Brigham, the fruits of whose labors in the 1920's led to themodern College Board tests. Alandmark study with implications forhigher education was a sweeping surveyby Learned and Wood of studentachievement throughout the Stateof

Pennsylvania in both lowc_sr andhigher institutions ofeducation.51 In the late 1920's, these investigators(in a mammoth project supported by the Carnegie Fund for theAdvancement of Teaching) looked over many institutions anddepartments with before-and-after measuresof academic achievement. The findingsobviously took the investigator's

breath awa'y momentarily.In the main, far more variationin level of growth than had been anticipated wasfound both among institutions and among departments withinthose institutions.(Indeed, college sophomores in some departments werefound to rank with tenth-grade

students in the general population!) Needless to say, the findings were as threatening asthey were informative; and today,with increased knowledge of the limitations of testsand of their reliance on thecurricu- lum, a less dramatic case wouldbe made than that of theinvestigators in their report.But our first finding of greatsignificance for a measure- ment science with implicationsfor higher education is thatconsiderable 72 diversity exists among andwithin institutions in the level ofinput and output on studentachievement.

One might suspect that after suchfindings had been reported, other states, or collections ofinstitutions, or guidance interests, would rush to cl-Art the levels ofdiversity among meaningful groups of institutions.But such was not the case.It may have been some part of the American dreamthat holds a college is a college is a college; it may have been that institutionsdo not willingly submit to the prospect of a ranking, or thatadministrators are concerned with not tipping internal or externalbalances of confident complacency; it may have been that officials ofmulti-institutional agencies with responsibility over many institutions did notread the findings, or, if they did, were afraid of threatening theinevitable half of the constitu- ent institutions that wouldfall below the average in any ranking; or, it may have been that colleges inthe depression years were so strug- gling for survival that those studentswith funds for tuition were happily accepted, whatever their credentials.But the next significant attempt to openly describe a collectionof institutions probably did not come until some 20 years after thepublication of the Pennsylvania study. In 1957 J. A. Davis, from theprotected position as staff member for the Board of Regents for the UriversitySystem of Georgia, published a "Counselors° Guide,"52 giving test data on entering freshmen in the seventeen-college system ofthat state. 73. the store The mostsignificant effort atdeveloping and expanding work at the Minne- of measures ofhuman traits andattributes came in Institute53 in _:ie 1930's. This sota EmploymentStabilization Research implications effort was notconcerned particularlywith measures with vocational for higher education,but rather withgenerally lower-level aptitudes that hadrelevance for vocationalsuccess.It did provide " samplings ofbehavior" someexperience in extendingthe test-contained search for constel- to a variety of newareas, laying abasis for a later lations of unitary "aptitudes" ; and it didbegin to signal thecomplexity of human traits andthe lack of precisenesspossible in defining the minimal levels necessaryfor an individual'ssuccessful performance in some societalrole.

in Shortly before thistime, Spearman inEngland and Thurstone ability testing the United Statesbegan to examineconventional mental could be divided into to determineif the variancethese tests captured ability repre- separate, unrelatedtraits, and if ourconcepts of mental worthwhile but ofdiffer- sented aconglomeration of thingswhich were exten- ential applicability.The tool they usedin this process was an sion of correlationaltechniqus that came tobe known as factor whether analysis. Althoughdebates raged forseveral decades as to making there were general(Spearman) or specific(Thurstone) " factprs" perfection of the up mentalability, the controversyled more toward theory was statistical techniquethan towarddemonstrating that one 74 more nearly correctthan the other.Perhaps the most significant aspect of the controversy wasthat the argument tended to hinge on logical or mathematical proofs, ratherthan on empirical or utilitarian ones. After a littletime that permitted obeisance to suchintellectual niceties, the hue and cry began to emergefor attention to criterion measures (those which maybe used to validate a predictor measure, or reveal, throughexperience, what the initial predictor measure was all about) and to empiricism in testconstruction (wherein the search is for expedients that work, whatever thetheoretical basis for defining a trait). With the adventof the computer age in the early1960's, factor analysis itself was to become a tool for testrefinement or clean-up of items, and an indicator of tracesof new factors that could be amplified to form new tests.

As the country began to emerge from thedepression toward the end of the thirties, there were someextensions of measurement to edu- cational arenas that involved morethan the conventional ability and achievement tests.In several instances, new colleges and adesire to innovate carried a socialscientist along. This was the case at Bennington,54 where the research model provided for a hypothesisof attitude change as well as cognitive growth.C. R.Pace55 picked as a central problem thestudy of the institution throughwhat could be observed in the subsequent lives ofits graduates, as did Chamberlin et al.,56 but the measurement aspects of these studies Were not highly As noted in the first part of this paper,World War II provided some flesh on thebones of Professor Brigham's PsychometricLabora- tory at Princeton. By theend of the war there were some onehundred test technicians and statisticiansin residence, allowing the labora- tory, as the Educational TestingService, to be the agent for the three adopting organizations (the CollegeBoard, the American Council on Education, andthe Carnegie Foundation for theAdvancement of Teaching). That agent, with its experience gainedduring the war years, provided some measurementdevices to what was probably the next landmark event in measurement inhigher education: the Coopera- tive Study of Evaluation in GeneralEducation,57 sponsored by the Carnegie Corporation and the AmericanCouncil on Education, and directed by Paul Dressel at MichiganState College. This study was notable because it began with elaborate andpainstaking efforts to define teaching objectives in generaleducation and in a number of areas therein.It paired test or statistical technicianswith subject matter specialists, and proceededthrough the specific and tailored construction of measuring devices to samplegrowth among students toward these teaching objectives.It also applied those measures in a before-and-after fashion with anexamination of institutional or instructional differences among or withinthe nineteen institutions

involved. = -

76

In retrospect, as we recall the boomin self-studies from accredit- ing commission activities as well as thestudy by the American Council of Education, the decade of the fifties can be seen as onein which measurement-oriented people were concerned with institutionalobjec- tives specified in some form amenable toobservation of student progress toward those objectives.In 1954 B. S. Bloom and his co- workers published a Taxonomy of EducationalObiectives,58 which focused on the general cognitive domainsalthough this landmark effort has had more direct use in measurementconstruction in other countries than in the United States.In this country there seemed to emerge an awareness of thegulf between what one might call general objectives of higher education, as specified in thegoals section of a college cataloglie(and even as purified by the measurement-oriented person), and the now fairly routinized procedure for building asubject- matter achievement test to reflect thehighly specific set of tasks implicit in a course unit within a discipline.

Drawing principally on the burgeoning interest inclinical and/or counseling psychology and a resulting interest in mentalhealth set off by Carl Rogers in the forties, a new entrant inthe area of measure- ment was to emerge and develop. This wasthe test of personality, important in this account for the negative role it was toplay in the fifties and sixties in studies of college students. A number of entries, ranging from pathology-oriented instruments such as theMinnesota

, 77

Multiphasic PersonalityInventoryqg to the normal psychology Guilford-

Zimmerman Temperament Survey,60were developed withclinical or normal populations of college students or adults,and were to be adapted, as will shortly be seen, to studiesof the impact of college.

Thus, in the first half of the twentieth centurythere had emerged some reasonable experiencewith a variety of tests of general scholas- tic aptitude, and some guidelines which madethe construction of a reasonably good test of acquisition of factual knowledge in agiven area a relatively routine proposition.Application of these tests beyond the college admissions situation, or in multi-collegestudies, was relatively rare.There.were a few attempts to measure by sorting indi- viduals into outcome categories, as in the several follow-upstudies. An occasional social psychologist concerned himself withattitudinal change as the function of a particular class of college experience rather than of a laboratory kind of treatment. Some questionabletools for exploring the personal impact of college grew out of themental hygiene movement. But for the most part, psychologists andstatis- ticians learned how to purify tests toward the prediction of grades in college; colleges learned that these could be usefultools in selec- tive admissions situations, and not much happenedbeyond that.

The decade of the sixties It is not inaccurate to generalize that measurement specialists 78 were, by factorsintrinsic in their role,implementers rather than innovators in their firstfifty years.This staff rather than line func- tion was to be changedrather sharply by a significantchallenge from outside the ranks.The start of a completely new eraof appli- cation of measurement,and new role's for the measurementscientist, probably began with P. E.Jacob's study, 61 published in1957, of value change as a functionof college experienca. Hisfinding that not much change takesplace threatened the very coreof academia, and many rose to the challenge: someto attack his findings,others to see for themselves.

If Jacob conceived the newmood, Nevitt Sanford(most precisely a personalitytheorist), through a collection of essaysentitled The Although AmericanColleg62e, . served as midwife for the birth. focusing frequently on relatively narrowpersonality areas of student development, he also struck provocativelyat many of the cherished beliefs and values of higher education.Nourishing this infant was the explosion of the college-agepopulation and a favorable economic situation, both of which encouragedthe prospect of rapid change in positive directions for almost allinstitutions of higher education. With such a start, the expansionof federal and foundation support for the technological revolution ineducation, and the beginning evolution therefrom, served to makeinvolvement in research not just possible but mandatory. Theresult is that at least some ten 79 studies, of as much significance asany ten in theentire past history of higher education,have just been published or arein prepara- tion. These include asubstantial volume by A. W.Astin63 on the college environment, provokedby the growing interest ineducational climate and based on well-designedcross-institutional studies by the American Council on Education; awell-financed review of the literature on impact of college now in progressat the Survey Research Center of the University ofMichigan;64 and a provocative study of eighteen liberal arts colleges nowbeing concluded by MorrisKeeton.66

This remarkable acceleration ofgeneral interest in researching, through measurement-orientedstudies, has involved both the measure- ment specialist and thegeneral social science faculty,who now use measurement and statistics asroutine tools.In effect, it has been a migration of thegeneral social scientist into his owncollege back- yard. Thus, Clark andTrow66 have focused on student subcultures as othersociologists have examined societalsubcultures.Sanford67 concerned himself with personalitydevelopment as a function ofthe college educational experience in muchthe same manner that other psychologists have sought meaningfulrelationships between personal- ity and early childhood experiences.On the other hand, there has been a rapid growth of a newphenomenon within the college adminis- tration itself:this is the institutional researchoffice or person designated to conduct studies related tothe maintenance and further 80 development of theinstitution.

In a recent paperdirected most specifially to this latter group, H. S.Dyer68 has asked the question:"Can institutionalresearch lead to a science ofinstitutions?"The question,particu'llarly as it

tt comes from sosophisticated an observer, , impliesthat we do not yet have a science ofinstitutions, that the socialscientist has not yet achieved it, and thatthe new institutional researcherwhocoulc conceivably contribute toitsdevelopment may not beable to carry it off.Dyer's argument isthat the institutionalresearcher may achieve a science of institutions, if he can integratehis views with those of the discipline-oriented social scientist(i.e., if he can

take on the values,strategies, andfools of the scientist) without losing his focuson the missionof the institution, andif measurement is permitted to play acentral role.

To those personssteeped in the liberal artof higher education, these comments may seempatently trivial.Social scientists, reasnable enough (orunreasonable enough) people atfaculty meet- ings, seemimpacted with technologyand with specificitywhen they are involvedthrough their disciplinewith an educationalproblem. Institutional researchers arealso pleasant enough asindividuals, but are frequently viewedby faculty educationalphilosophers as market researchers at best,tools at worst, of the newmanagement 81 mainstream in highereducation. Measurementreeks of standardized tests.If these indeed arethe forces and thebasic ingredient of a new scienceof institutions, can onereally expect much to comeof it? Considering thetruly basic problems--e.g.,contrivance of genuine academicfreedom, the best synthesisof the conscience vs. intellectquestion, manipulationof the individual andthe environ- ment to assure adeepening of intellectual awareness,or the future of predominantly Negrocolleges--what can one expect to see achieved from the me_surementscience school or theinstitutional research school?

The purpose of the secondpart of this reviewis to explore the state of thedevelopment of a science ofinstitutions of higher educa- tion--how far it has come,where it has succeeded,where it has failed; and, what directions areindicated for the future.In this review, Dyer's notionthat measurement iscentral will be taken seriously; this not onlyhelps to define and delimitthe area of inquiry, but also may indeed, asDyer suggests, form thecrucial component. For measurement, asDyer uses it, is not aloose synonym for "test," but the endproduct of the process ofdefining some quality of concern with sufficientprecision that an investigator(or others) can make more exactcomparisons among individuals, groups,or instititions in regard to thatquality.In mOst instances, the measure will have limitations, for itusually involves abstracting partsof a 82 totality that in practicehas been fuzzy atbest.But the process of defining helps toclarify what the measureis not as well aswhat it is, at leastfor operational purposes.This brings someorder, control, and openness toinspection by others intothe situation. Also, the conversion,through measurement,of a quality to ametric unit permits evaluativecomparisons of productsmeasured. The base quality can thenbe studied systematicallyin relation to other qualities, and studies canbe replicated towarddetermining how the quality develops andfunctions. Although inthe beginning of a science scope issacrificed to precision,that precision becomes the base for building.

We also need atthis point to define"research," for thistitle has been given toa.wide range of activitiesfrom head-counting or in- gentle speculationfrom a few observationsto massive projects volving elaboratestatistical treatment ofdata. Research isthe systematic inquiry intoconditions bearing oncertain events or out- it operates comes.It starts with one or morespecific questions, and through a preconceivedand deliberate plan,designed to identify and properly control relevantvariables so that their truemeaning may be better understood.The deliberate planinvolves procedures designed to sharpen ormake objective thefocus on the variables of concern, and toexclude biases, manyof which are likely tobe subtle, which may lead toerroneousinterpretations or conclusions. Many who have attemptedto support and encourageevaluative research in higher educationhave attributed such a status toalmost any generalizationfrom experience orobservation, or to any honest inquiry whether structured ornot. Thus, W. H.Cowley69 sees its beginnings in American highereducation in 1701, whenIncrease IViather, then president ofHarvard, functioned as aneducational consultant to the foundersof Yale. Yet true researchmust stand on the evidence it marshalsrather than on the statusof the individual making a pronouncement,the versatility of his argument, oreven the eventual proof that he wasright. For a long time, manyeducators bought the argument thattraining in Latin afforded aunique and valu- able mental discipline forthe learner.. Yet as othersubjects appli- cable to new societal needsbegan to vie with 'A in thecurriculum, and when true research onthe impact of training inLatin was conducted (i.e., when some progress inmeasuring ability and impactof mental discipline was achieved), it wasgenerally found that the superiority of students with Latin training, orwho attended institutionsrequiring healthy portions of Latin inthe curriculum, could moreaccurately be ascribed to the ability of thoseattracted to or sorted into coursesin Latin than to any habitsformed*as a consequence ofclassic studies. Confident, even experienced,opinion can be quitemisleading: in higher education it tends tooversimplify (e.g., "goodteaching is the crucial force in intellectualdevelopment"), or to be too uncritically

:7:77 4:11 84.

high acceptant of utilitarian orstatus values(e.g., "a college with ability enteringfreshmen is aqualityinstitution").

of the Even within such abrief historicaloverview, as this one examine the period from 1900 to1960, it isinteresting to attempt-to of forces at work which mayhave hampered anearlier development the current decade. the researchliterature that hasbegun to appear in acquiring com- The major deterrentwould seem to bethe difficulty in for parable data across a rangeof institution:J.Two factors account universities tend tobe this.First, the fact isthat colleges and independent entities, orif parts of a system,seldom have anadminis- trative head whowould considermonitoring, throughresearch, the of diversity, separate componentsof the system.Second, the very fact and the fact that evensuch an index asadmissions test score averages overtoneS for those insti- for entering freshmenhas terribly threatening tutions below the top,indicates that measurementresearch across institutions involves somevery sensitiveareas.

The current decadehas seen a newforce as a mostimportant development toward ameasurement-based scienceof higher education: this is the researchorganization or team,strategically or by happen- financial stance situatedwith access acrossinstitutions, with ample resources forcostly inquiries orsustained effort overtime, whatever the vagaries of funding,and frequentlywith some practicaloperational 85 channel permitting research data as aby-product. The Center for the Study of Higher Education atBerkeley, started in 1957, the Center for Studies in Higher Educationat the University of Michigan, established in 1955, and the Institutefor the Higher Education at Teachers College, Columbia, startedin 1962, are the early pioneers of university-based centers. The regionaleducational boards, par- i.:ularly the Western Interstate Council onHigher Education and the Southern Regional Education Board,have provided themselves with a high springboard andhave taken a magnificent dive. Among the testing organizations, the NationalMerit Scholarship Corporation, under Holland's leadership, gave an earlyand pervasive focus to cross-institutional measurement research.Holland carried this with him when he moved to the American CollegeTesting Program in 1959. Although the, College Board and Educational TestingService, its research and operational arm since 1947, havefrom their beginnings looked at problems of general interest inthe administration of certain functions within institutions, the year 1963 sawthe establishment at ETS of a research group concernedbroadly with higher education, and in 1964 a new program to providecolleges, at their option, a variety of measurement instruments, packagedresearch designs, and a variety of data processing services.In 1964 the American Council on Education, a leader for many yearsin seeing studies initiated, acquired its own in-house research program. 86

That these giants havebecome the leadersin one area of concern is no accident. Whenthe chairman of a governing body of a system of colleges, with budgetarycontrol, demands cross-institutional studies, the fearful memberinstitutionshave found ways to dodge or defeat the attempt.Cross-institutional studies can, as hasbeen noted, be threatening, for whatcollege does not have a faculty zealous of its own particular independence and brand of control, and a few skeletons inits closets? One needs an organization with easy access to manyinstitutions, apast or a base without theblemish of former crusades, and a financial base to sustain costly activity.

The reader, then,should be aware that in the remainderof this review we are indeed looking at a young plant justready to sprout; we cannotconfidently predict how tall it may grow, orwhether it will prove tobe prickle or pear.In any event, cultivation may be essential ifit is to thrive; or, on the otherhand, it may turn out to be a rampant weed that will take someunusual efforts to stamp out.

The Enlar ed Concern with Student In u If any fact about higher education iswell-established it is that a measure of pastperformance (scholasticachievement), together with a reasonably good test of mentalability (scholastic aptitude), is a good indicator of what gradethe student will achieve in college. This finding has held over the years and overthe range of institutions, 87 from vocationally orientedjunior colleges to selectiveIvy League institutions.

It was pointed out inPart I that the first exhaustivereviews of such studiesbyliarris70 supported fairly well the conclusionsof Fishman andPasanella71 two decades later.Performance in secondary school (as attested throughrank in class or grade-pointaverage) has been again and again demonstrated asthe best predictor of college performance. Adding a test score tothis measure improves the pre- dictability, probably because, asJ. R.Hills72 has speculated, it helps correct for differential standards amongsecondary schools in grading practices as well asbecause it reflects the basicmental equipment needed tounderstand and retain academicmaterial. Other measures--of interest, personality,attitude, motivation, or work habits--have not improved theprediction, either because the measures of these qualities thathave been contrived arefaulty, because these qualities are alreadysubsumed in the measure of pastperformance, or because thesequalities are not so uniformlyand critically impor- tant as ability andachievement.

Most of the several thousandstudies reported in the research literature involve single institutionsand the search for the most efficient weightings of the severalingredients of the predictive combination, or the (generallyfruitless) search for ways to improve 88 that combination.In the weightingstudies, minor variationsthat make sense have beenfound:for example, engineeringschools find that some minorimprovement is effected byweighting mathe- matical components moreheavily than would provebest for liberal arts colleges.But these variations arerelatively inconsequential.

Much more substantial are thedifferences in level (not in content, relationshipwith grades, or weighting) onindices of academic promise among institutions ofhigher education. This was apparentin the classic Pennsylvaniastudy of Learned andWood;73 the most recent major survey isthat by J. G.Darley.74 No formal and ex- haustive studies of why thesedifferences exist are known,although Astin75 has studied selectivity against a varietyof other descriptive indices. His data show that thedegree to which an institution must he or can be or wants tobe selective, the severity ofthis selection (a function of number of applicants vs.number of places made avail- able), and the period of time theinstitution has been selectivewould appear to be the mostpowerful influences.Apparently, the practice of exclusion, together with thefocus in selection on arather par- ticular and unitary kind of variable(academic promise from tests and past record), broadcasts ,topotential applicants somerather powerful signals which are strong enough to encouragesome to tune in on other frequencies. The age of theinstitution also appears to play some part, particularlywhere the history of the institutionextends 89 back more than a century. This may be a function of the relationship between socioeconomic status and intelligence, the fact that in earlier days going to college was confined to narrower bands of ability at the top of the distribution, and some sort of enduring bond between a socioeconomic class group or strata and the institution. Institutional commitment to broad population groups appears also to play a part.Public colleges supported by broad-based tax funds have been generally prone to build to accommodate additional numbers when demand from prospective students among its constituents increases (although there are differences among state universities, whether by age, availability of other kinds of institutions with differ- ent selection policies, or from a built-in branching system as in California today). Another powerful influence appears to be the dominant curriculum of the institution, or its designation as an insti- tution for training some particular occupational groups. The hierarchy here runs from institutions which are training for the learned and scientific professions, through colleges tuned to the band of middle- class business and professional service occupations, to teacher training institutions and to junior colleges that are primarily devoted to vocational specialities or remedial work. Institutions that have been restricted to Negroes, who for one or another reason tend to score at the bottom of academic promise indices, tend to produce the lowest mean scores for entering freshmen (or, for that matter, 90 for graduatingseniors).

That these differences existis pervasive knowledge.However, the magnitude of the differencesis not so well known.In most states, there are institutionswhere a student in the top 10 percent of his class would fall inthe bottom 10 percent of anotherinstitution within that state.In the College Board metric,institutions may be found whose entering freshmen average275 or 725; in terms of the , conventional I.Q., the range wouldbe from the low 80's to thehigh

140's.

What impact these 'differenceshave for a more consciouscommit- ment by given institutions tospecific talent-manpower levelsis

uncertain.Thorndike andHagen76 have charted talent ranges (or, (. in our context, academic promiseranges) for a wide spectrum of occupations, but these data, aswell as an earlier observationby D. E.Super77 and the still earlier data on draftees in World War11,78 show that each occupation has arelatively wide band of abilitywithin it, and that the relationshipbetween occupational field andmental ability does not so much stemfrom a relationship betweenability and occupational success, but ratherfrom a relationship betweenability and sustained study beyondhigh school (together with thefact of educational requirements, meaningful ornot, for some occupational fields). The picture is further complicatedby findings such as those study of of JamesDavis,79 who provided evidence,in his follow-up college graduates from anumber of institutions, that thebrightest students in each institutiongenerally majored in science, butthat institutions, sciene majorsrepresented a wide , taking the totality of range similar tothe institutional differencesthemselves. At this point, and probably atfuture points,one cannot saywith any assurance what minimum abilitylevels areneeded for different manpowerspeciali- ties. Accountantsand doctorsas groupshave higher mentalability averages thanfile clerks, who do better on teststhan lumberjacks; but some lumberjacksbecome doctors,and some doctorsmight make better lumberjacks. What can be said mostprecisely is that given the varying dependence amongoccupational fields onacademic content, together with thetraditional teaching methodsinvolving symbols

(verbal or numerical) and abstractions, therather fuzzy hierarchies will probably continue.We need to expand theinventory of abilities and determine their preciseoccupational relevancebefore stating that aiven training facilityshould admit students ingiven ability thus far rangs. Yet,efforts with expandedaptitude or ability tests

haye given noindication that this ispossible.It is more likely that men will neverdiscover nor agree on aprecise constellation ofattri- butes for any given field,but they must make somecontinuously revised decisions as to thestyle and content oftraining. As a social animal, man needs enoughcommunality to communicatewith others, 92

particularly those with whom he works; but what is needed in a given field that one worker cannot provide, another colleague will. The point is this:there is no evidence from measurement studies that .establishing training facilities to cater to certain talent ranges, and manipulation of the size of these facilities to control flows of trained manpower, is a reasonable possibility.Rather, it would seem that one should start from the other end. What content orskills (including academic or mental skills) are needed, and what alternate training procedures are likely to bring students, among whatever levels avail- able, to mastery of that content? What diversities does an occupational field require? In this work, measurement is more likely to be useful as a tool to define proficiencies and to attest to when theyhave been reached than to answer completely the question of who shall be trained in what. The more precise proficiency criteria may lea.d to the develop- ment of more precise promise criteria. At the moment, we have a general promise criterion for a general academic proficiency criterion. The'faster the learning pace, the more abstraCt it is, and the more memorizing of abstractions, facts, or images is required; the more reading or problem solving involved, the more complex the subject matter; and the more traditional the system of instruction, the greater the level of traditional mental ability that is required.

It has been stated that the general as well as the academic public perceives a diversity and a hierarchy (though dimly) of institutions in 9 3 terms of levels of mental ability ofentering students, and that this hierarchy does indeed exist. Another perception amongthe general public and most of academia is that the quality of aninstitution is ,synonymous with--or, at least,highly related to--the quality of entering freshmen.It is reasonable to assume that the higher the mental ability level of students, the faster is the pace that canbe set, the richer is the content that can be required,and the greater the levels of competition that can be nourished--all of whichimply a quality capability. But reasonwould tell us that after rather than before measures are needed (or better yet, both before-and-after measures) to determine what an institution has done for or to its students. Using admissions prerequisites as standards is a dean's way (instead of a facultyway) of controllingor attempting to control-- a faculty, who if they failed all studentswould have nothing much left to do.This logic suggests that our perception of the greater quality of Ivy League' vs. struggling church-related college, for exaMple, is more rationalizational than rational.

Added to the fact that it is possible to take good student raw material and do nothing much with it, thus casting doubt on thisinput characteristic as an infallible quality criterion, there isthe differen- tial attrition rate among institutions. From occasionaldescriptive reports or records, it is apparent that some institutionsfail only 2 or 3 percent of their students, while others may drop 50 percent or more 94. before the sophomore year.In other words, grading standards vary and are another traditional wayof attempting to maintain andcontrol the quality of the institution or program.Yet, no competent studies across institutions areknown that shed much light on theforces 'behind differential attrition standards.

Generally speaking, records show thathighly selective institu- tions at one extreme, and institutionshungry for students at whatever cost at the other, tend tofail relatively small proportions ofenrolled students. Davis° data on studentsmajoring insciencen would indi- cate that attrition "standards" are a function of the institutionrather than of absolute performance levels orfield-related standards. Studies within single institutionsexperiencing rapid change in the level of entering students over a period of afew years (e.g., the studies by S. C.Webb81 at Emory and Aiken82 at the University'of North Carolina) show that faculty members tend to maintainwhat they perceive as a going rate in assigning Pi's, whateverthe fluctuations from year to year in ability.This and foregoing observations suggestthat attrition rates are basic to the institutionrather than to the student or major field, and that the need to retain students or abelief in the infinite superiority of the student are factorsthat depress attrition rates.It would seem that some importantand intriguing studies could be con- ducted which might seek out faculty,administration, board of control, or constituency or manpowerfactors which affect the establishment 9 5 of various attrition standards.Does overcrowding by anadmissions office press faculty intothinning the ranks, or does itsignal that more students are tobe given the higher educationtreatment? What part is played by manpowerneeds? Has the acceptance ofthe Negro in a wide variety ofhigh-level jobs put pressures onpredominantly Negro institutions to graduatelarger proportions of enteringfreshmen?

Crucial in cross-institutionalstudies of the sort that mightdeal with such questions is thefurther development ofproficiency or academic attainment measures.Davis83 has argued for a national grading system, althoughsuch would be quite threatening to many colleges (given the relationshipbetween academic ability and per- formance, and the diversity ofability ranges overinstitutions), and it is difficult to imagine howsuch a system could beestablished. Another problem is the commonvariance--or overlap--between tests of ability and tests ofachievement (for example, theNational Teacher Examinations are more ability testsin disguise than tests covering content of teacher preparatory programsor of the skillsuseful in teaching).

It was noted earlier thatpersonality tests grew principally out of the early vocationalguidance tnterests and the mentalhygiene movement of the 1940's.It was also noted that testsof personality have been used, but haveyielded little, in studies ofentering freshmen. 96

This refers, mostexactly, to the use ofpersonality measures in attempts at predictionof academic performance.

In a search fornon-cognitive factorssuch as motivation, per- severance, stability,and the like which mayaffectacademic per- formance, a virtual legionof investigators havetried each test as it comes out; there are morethan a hundred studiesreported, and probably many more unreported,that have examined therelationship of the Minnesota Multiphasic Personality Inventoryto grades. Such work may be summarizedby the statement thatthe normal and abnormal psychologists concerned withhuman development havetheir theories and criteria, and theacademicians concerned witheducational develop- ment have theirs, and never,it seems, would the twainmeet, An early attempt to break thisdeadlock was made by Fricke atthe Uni- versity of Michigan, whosought responses to a somewhatwild array of items (e.g., preferencefor poodles or Germanpolice dogs), and then observed if particular responseshad implications for later academic performance. Hisstudies, reported in theresulting test manual, 84 have generallyproved of more use in Fricke'shands than in the hands of others atother institutions, a not uncommonoccurrence. Another relatively significantattempt was the constructionof the

Omnibus PersonalityInventory85 by a team at the (then) Center for the Study of Higher Education atBerkeley, where the concern was ex- pressly for contriving a set of measuresuseful for exploring problems 97. in higher educational development. This instrument still reeks heavily of the clinical heritage, for it was built by selecting items or scales from existing personality inventories that were felt to be relevant.

Potentially more valuable instruments were the Edwards Personal Preference Schedule86 and the Student Activities Index.87 Both of these instruments were drawn from the fifteen normal needs defined by Murray in 1938,88 and seem to have fairly direct relevance to the efficiency of students' functioning in the academic situation. The Student Activities Index was a parallel instmment for another (the College Characteristics Index)89 established specifically to study how the press of the college environinent might serve the needs of students or frustrate them. Although these devices may have been carelessly or casually used in most of the studies conducted with them, not a great deal has come from either in defining important differences among entering student bodies.

Worthwhile results over a range of institutions have been reported by G. E. Schlesser90with his Personal Values Inventory, a personality meFisure designed specifically for predicting grades in college. How- ever, not many of these reports are readily available in the literature, and the fragments one finds here or there imply that Schlesser's failure to include high school performance in his studies (arguing that the 98

PVI is a substitute for the uneven grading systemsfrom school to school)may mean that the PVIis, in effect, a self-report of high school performance.

One of the more promising leads, with regardtomeasures of personality applied to the student input situation,has come from several investigators who have looked at thedifferential personality profiles of groups of students applying to various kindsof colleges. Several studies by the National Merit ScholarshipCorporation's researchteam,91 andsome atBerkeley,92 have shown that there are reasonable differences in such areas as achievement motivationwhen one looks at high-abilitystudents who select competitive as opposed to non-competitive colleges. Thismay suggest that powerful college image factors perform a kind of sorting function, andthat with the restriction of range on the more elusive personality factors their import for achievement or development is clouded in studiesof single institutions. To complicate matters even further, it isreasonable to assume that a given personality trait--e.g.,introversion--could facilitate achievement in one kind of program or institutionand retard it in another.

The way out of these dilemmas would seem toinvolve, first, a more searching attempt todefine personality constructs relevant to the academic demands and to learning(e. g., who has produced a 99 good measure of interest in ideas?), and then to seethese applied over a variety of institutionsin conjunction with specific analyses of their image and their challenges. The diffeiencesin mood, if you will, in student bodies at Antioch or Ball State are apparent, and although the college atmosphere plays some small part it may be expected that image and input play a larger one in shaping it.

Although their instrumentation is not yet ready for general testing and application, it would be amiss not to cite the tack being taken by a measurement-oriented group of personality theorists, centering most exactly onMessick93 at ETS. This group started with a concern for studying the error that seemed to systematically affect the per- formance of certain individuals on tests; for example, it was found that tests with items answerable in yes-no fashion were affected by a trait labeled "acquiescence"--acquiescentindividuals tended as a matter of course to prefer a "yes" response,while the non- acquiescent tended to favor a "no."After some brief effort to estab-

lish tormats that would not be'biased by some such response styles, Messick and his associates began to look at the styles as important human traits in their own right. What is emerging from this study is an inventory of problem-solving styles, related to the "cog- nitive styles" of other investigators, and some procedures for deter- mining their relevance in a variety of situations. The general tenor of the developing theory is that different disciplines, or learning 100 tasks, may involve or favor different problem-solvingstyles; under- standing how these processes may function could affordtalent identi- fication and selection devices, but could better provide abasis for .diagnosis and specific training in the appropriate styles needed. Although this may be the measurement scientist's way of legitimizing the popular belief that successful mental activity in mathematics is not the same as that in English Literature, there maybe promise for some important breakthroughs in measurementand in the psychology of learning.

A new and relatively rich area of studies of student input have focused on relatively simple biographical kinds of factors--the student's socioeconomic class, the nature of his family and school experiences, and hiS attitudes on educationally relevant issues. The crucial modern work started with Clark andTrow,94 who attempted to define student subcultures from biographical factors reflecting their purpose in going to college. These efforts led tothe construction of the College StudentQuertionnaires95 by Peterson, who has reported some of the differences observed over a rangeof institutions in their entering freshmen.96 Within the past two or three years,Astin97 and the American College Testing Program researchstaff98 have also assembled some information on how the entering students vary in background, attitudes, and aspirations. Attempts to develop scales 101 of such items reflectingpsychological or developmental traits are still in very primitive form(e.g., Peterson has brief scales of such traits as family independence,cultural sophistication, and the like); but this may improve soon. Thecrucial aspec-t-of the matter is that there has been strong interest amonginstitutions in looking at their student bodies in such anintensive fashion. Knowing one's students better, particularly in these timesof rapid change in college attendance, would seem a promising basis forbetter-directed educa- tional treatment.

In regard, then, to an examinationof measures of student input, the last decade has seen a moreopen.recognition of the diversity in intellectual ability that marks American institutionsof higher educa- tion.But more important, we are beginning torecognize that a host of other characteristics may haverelevance for the educational or institutional mission, and we are beginning tobe less concerned with seeking a preconceived qualityof academic aptitude (defined unidimensionally) and more concerned with understanding someof the more subtle variations among studentsattracted to one or another kind of college. A formal examinationof the attitudes of entering students may also provide a betterbasis, eventually, for assessing change in areas beyond the reach ofhidebound achievement tests. But of greater importance, we may be justentering an era when 10 2 training can be tailored tothe student, ratherthan reserved for those who learn conventionalthings in conventional ways.

The Enlarged Concernwith the EducationalContext The measurement mainstreamin highereducation-. grew (some would say to floodstage) with a focus on the measurementof human traits and qualities throughtests designed to tap someunderlying continua. Whatever the specificgoals of any institution ofhigher education, it must have as itsmajor function the guidingand enhance- ment of individualgrowth and development.Measurement tools for such qualities as learningreadiness, scholastic aptitudeand achieve- ment, learning styles,and the like have obviousutility for managing and monitoring some basic concernsof colleges and universities. Indeed, the means anddispersions of scores of studentpopulations

may expressimportant institutional qualities.

However, having such measuresand a background of social science research strategies, a personconcerned with institutional functioning should now seek some waysof measuring institutional qualities with the aim of findingassociations between whatthe institution is or does on theone hand and what happens tostudents on the other.Yet, until the last ten years,not much happened in the way of defining dimensionsof institutional diversitybased on instibitional qualities, or in constructingscales to reflect these 103 diversities.

Several factors may account for this surprising omission. Typological categories formed on such bases as control, type of curriculum, size, and the like have been used for some time, and serve identification and record purposes.There are popular tendencies, even among professionals, to view colleges and universities along a unitary continuum of goodness-badness (although simple reflection will indicate that an institution must have many different qualities of goodness and badness, or that institutional qualities good for some students may be bad for others). Most measurement specialists cut their professional teeth on the study of differences among individuals, of statistics and methodology, or both.Statistical procedures such as analysis of variance permit one to look at the possible influences of different situations on the student,thus reducing the necessity of having continua of institutional qualities.Persons concerned with the theory and philosophy of higher education have, on the other hand, seldom come from the disciplines of statistics and methodology or of measurement, and if they do they are frequently pressed in their professional roles to grapple with urgencies of housekeeping such as funding, arbitrating among various proponents of curricular change, and the like, leaving little time for such a staff function as measurement- based research. 104 For some time there have been two basic kinds of approaches to institutional analysis and assessment that are relevant to what we have called here the educational context. The first is essen- tially the case-study method, conducted by one or several perceptive observers with some variety of experience over a number of institutions. A good example of the results of this method is the set of ten studies provided by David Boroff in Campus, USA.99 Measurement as such plays only the slightest of roles in this approach, and the consumer is at the mercy of the ability, wisdom, and experience of the observer.

Another time-honored approach to studies of the educational context in institutions of higher education is that typically employed by accrediting commissions. This may include case studies by the natives or visiting specialists, but it also hinges on the collection of descriptive data on characteristics assumed to be necessary com- ponents of a favorable learning environment. A library is judged to be an essential component; hence, the number of volumes, circulation rates, or the budget for continuing acquisition form indices that may be contrasted among an array of institutions, and used as a basis for qualitative judgment and standards.

Although this procedure grapples with some subtletieS, it has three serious flaws. The first is that it must be based on the norms of current realities. Who is to say, in any absolute terms, the 105 number of library volumes nece3sarfor a gi-ren discipline or insti- tution, other than someone with one eye on tha goingrates? With the advent of the computer as an instructionaland research tool, shold ii riot be incorporated into the minimum educationalessentials? Can it be so incorporated until more institutions acquireone? What a college requires to fulfill itsfunction should be the result of a careful series of studies to determine exactly what can be donewith different resources. One remembers too well Anne Roe°s100 finding that the small midwestern college, typically without extensive labora- tory equipment, is the major spawning ground for eminentphysical scientists, or that Russian schools were able to accomplish a great deal with homemade laboratory apparatus.

A second serious flaw, not unrelated to the first, is that what an institution has, and what an institutiondoes with what it has, may be two different things. There are those persistent critics who point out from time to time that this or thatexemplary educational facility houses a complacent faculty and a thriving country club ofstudents.

A third flaw frequently inherent in this approach has todo with the complexities of qualities and goals ofinstitutions of higher educa- tion vs. the fact that what is generally observed comesfrom things readily observed or counted. We can, without much difficulty, count the number of Ph.D.'s on the faculty, or student contact hours; we 10 6 can contriveratios that seem also tohave some value.But the point is that too frequentlythis approach has taken asbase data the things that are available andthat a registrar's clerk orpresident's administra- tive assistant canassemble. The lesson from measurementresearch is that one must firstgrapple with a definition ofessential qualities, then seek ways of measuringthem, and then test their truemeaning through systematic analysisagainst other criteria.

An important adaptation ofthis approach toward attempting to systemize procedures, defininggroup-related characteristics, and studying their meaning grew out ofthe work of a team of measurement specialists at the National MeritScholarship Corporation. This effort, reported by Astin in 1962,101 involved taking some thirty con- ventional kinds of indices for a groupof more than three hundred colleges, and using factor-analytictechniques to determine the basic underlying dimensions of diversitytherein.This study reveled, in Astin's interpretation, six principaldimensions: affluence (wealth), size, public vs. private control,masculinity vs. femininity, realistic or technicalemphasis, and homogeneity. Thelargest proportion of the variances among these institutionshad to do with the affluence factor, which was made up ofsuch indices as measures of thecollege's financial resources, ability of students,proportion of Ph.D.'s on the faculty, and so on.

, 45' rf, C re. 107

The procedure of factoranalysis is limited, of course,by the fact tharone can onlyfind variances emergingfrom the elements studied that were alreadycontained in the measuresgoing into the analysis. But the componentsof ("loadings on") a factorreveal relationships among ingredients aswell as provide a parsimonious or efficient wayof dealing with a host ofimperfectly related variables. J. M. Richards andothers at the American CollegeTesting Program have extended this treatment tothe juijor college(finding six factors or dimensionslabeled cultural affluence,technological specialization, size, age, transfer emphasis,and business orientation).102

Although this kind of beginning istremendously attractive to the measurement statistician atfirst, on second blush one must return to the possibledimensionality put into the system,and whether, as in the accrediting commissionapproach, one has chosen initial measures becausethey are important or simplybecause they are available. One might develop anefficient way of describing foodby taking a number of measures ofthe contents of a grucery storeand emerge with apackaging factor, a spoilagefactor, a wet-dry factor; but the practical purposes ofthe shopper make the originaldistinc- tions of fish vs. fowl moreimportant in provisioningfor Friday's dinner.Blind searches may stumble on someuseful leads, but a guided search may be moreeffective.In the latter case, it would seem we are lesslikely to end up with minormodifications of things 108 we alreadyknew.

The most significantadvance in measuring institutionalqualities (not in terms of what it has yetproduced, but in terms of where itis now leading manyinvestigators andtheorists) can be attributed to C. R. Pace and G. G. Stern,who developed in the latefifties an instrument called the CollegeCharacteristicsIndex.103 They reasoned that the important educationalforces in the learning environmentmight best be revealed through the eyesof students, and they collected a number of statements concerningqualities or conditions that students might react to as generally true or not truefor their campus. The content of these statementsreflected perceptions of the competitive scholastic pressures, thL, status ofthe instructors, the topics of free or informal discussion amongstudents, the emphasis on athletics, and the like. Application of theresulting instrument over a number of diverse campuses allowed theresearchers to determine what items reflected prevalent differences, andhcw these differences might be grouped into scales.

The next major development of thisapproach grew out of an honest split which shortly developed betweenPace and Stern.Stern, holding more to the interests of the personaland social psychologist, and heavily influenced by the need-pressmodel of H. A.Murray,104 turned the resulting scales on the CollegeCharacteristics Index 109 toward measures that seemed to correspondwith Murray's theoretical model and which were concerned with the totaldevelopmental needs of students and the parallel forces in the environmentrequired for theirsatisfaction.105 Pace, more concerned with institutional dimen- sions that might have useful meaning for thegeneral administrator or faculty member, used a portion of the items, applied factoranalysis to institutional means (rather than to individualstudent variance), and derived scales which he argues reflect true institutionalqualities. His instrument, called the College and UniversityEnvironment Scales,106yields scores on five dimensions:(1) practicality:the degree to which personal status and practical benefit areemphasized in the college environment; (2) community:-the degree to which the campus is friendly,, cohesive, andgroup-oriented; (3) awareness: the degree of emphasis on self-understanding andpersonal identity, a wide range of appreciations,and personal involvement with the problems of the world; (4) propriety:the degree to which politeness, protocol, and consideration are emphasized; and(5) scholarship: the degree to which competitively high academicachievement is evidenced with concern for scholarship and intellectualdiscipline and interest in knowledge and ideas.

The stage would now seem to be set for studies thatmight 110 attract many researchers to test forrelationships between individual growth and institutional characteristics.Perhaps partly because this development is s-) recent and becauseearly work in the develop- ment of instrumentation has todo, in essence, with tidying up the internal characteristics of items--or perhapsbecause the major initial use has been more byadministrative observers than by measurement researchers--not a great deal 'aas yet been reported un thevalidity, or tested meaning, of the scales.Some potential problems are apparent, however. One is that students aresimply not aware of many important features of thecollege or university--for example, facilitation of faculty research. Another is that there maybe some important omissions of content in the. totality of itemsthemselves, or that items get outdated rapidly, orthat it is difficult to find items that provide as fair a set of stimuli for, say, a small,predominantly Negro college in the South as for an Ivy League university.Still another is the fact that the items are more oriented toward the percep- tions of the student than toward his actual behavior(e.g., the distinction between "A lecture by an outstanding scientist would be poorly attended" and "I have attended a lecture by anoutstanding visiting scientist during the past term").

A somewhat different approach from that of Pace and Stern was

reported by Astin and HoHandl" in 1961. They reasoned (and not 111 entirely withoutevidence) that those thingswhich comprise the educationally relevant personalcharacteristics of student bodies, together with the studentsrelative emphases(through their majors) on different courseareas, can beused to constitute theenvironment. Their procedure was toassemble information on averageacademic ability, institutional size,, andthe proportion of studentsmajoring in departments groupedinto six different areas--thelatter kind of characteristic having been found tobe quite stable for institutions over time. Testsof this approach againstthe College Characteristics Index provided someevidence that student perceptionsof the environ- ment are not unrelated towho the students are whenthey enter college as well as towhere they place their majoracademic interests.

This procedure; called theEnvironmental AssessmentTechnique, could be employed by going todata of public record; thatis, its components were made upof variances that althoughalready known could perhaps be organized moreefficiently for purposes of institu- tional definition. At least onemajor validation studyin did explore the meanings of the scalesagainst the students' perceptionsof the effects that the individual collegeshave upon them. From one perspective, the Astin and Hollandstudies give some usefulinsights about meanings of selectivity andprogramatic foci; from another, one might say thatbrighter students spend more timestudying and 112 are more likely tochoose a humanities major than aneducation or business major, so ho-hum.Crucial for our purposes is the fact that some important newefforts were made; there was sufficient study of the results so that thelimitations of findings could be recognized and the search could bepressed further.

In a newer series of studies done aspart of a major new research program at the AmericanCouncil of Education,Astin109 has taken particular note of the perception vs.actual behavior question, and has developed some environmental assessmentscales based on what students say they actually do. These are nowbeing used in a systematic multi-college, multi-goal continuingstudy that most assuredly bears careful watching.

The most complex approach, and certainly the onewith the most- comprehensive attempt to build on sometheoretical model of institu- tional functioning, is growing out of a seriesof activities led by Earl McGrath at the Institute of Higher Education atTeachers College, Columbia University, and involving a highereducation research team at Educational Testing Service.Little more than an occasional working papernohas been released on this project as yet,but the effort is so potentially significant that it would be amiss not to attempt a summary of it here. 113

This effort, as defined byMcGrath's team, has been greatly influenced by John Gardner'stheory" of institutional functioning and self-renewal. Adopting a focus on"institutional vitality, " they have made a variety of attempts todefine its essential nature or natures and its components.These attempts have ranged from case studies of prototype institutionsto the systematic polling of educational researchers or educational leaders.Although the project's main concern centers on forces ineffective innovation and institutional continuance and survival, rather than onthe more orthodox kinds of dimensions of previousenvironmental assessment work, the developing studies are beginning toprovide some entry into the dynamic interrelationships amongeducational and adminis- trative forces. One by-product that may shortlybe available for wider use is an instrument that may be used tosupplement student perceptions of the environment with faculty perceptions,and Provide some additional dimensionsthat relate to academic freedom, recep- tivity to new ideas, etc.

The important thing about all of this work of thelast decade would seem to be that serious attempts are beingmade to define and measure significant social, personal, and educationalforces that may characterize institutions. The most importantrecognition of the decade may have been that of Pace, when,through his College 114 and University EnvironmentScales, he said, "Let's measurethe institution, not the students."There are those who mayforesee a simpleinstitutional evaluationfunction and expect a single good-bad dimension, thusquestioning the practicability(because of threat) of applications; yet asis so often the case, careful attentiveness to an institution, asto an individual,shall surely reveal interesting and attractivecomplexities, and provide leads for understanding and effectivemodification in desirabledirections. This more mature and acceptantattitude will surely prevail,carried on not only by themushrooming concerns witheducational technology and research, but also becauseit has its own intrinsicrewards.

The Measurement of impact ofColleges Upon Their Students Not much will be said here onthe measurement of thecollege's impact upon its students,partly because so little has beendone that concerns direct measurementof outcomes, and partlybecause of an excellent and exhaustivereview by Newcomb andFeldman112 that is now in preliminary draft.That study, which I believewill be the most important milestone inhigher education research since Sanford's 19 62 contribution, coverswork in progress as well as work reported, and proVides acomprehensive frame of reference for future studies.

The problem in studying collegeimpact is that it is extremely 115 difficult to contrive a"clean" research design.One needs before- and-after measures, some waysof controlling personal(as opposed to context) factors, some wayto separate the impactof the times from the impact of theinstitution, and some way to separatesimple maturational effects fromthose produced by contrivededucational intervention. Another problemis how impact will bedefined and measured. Simple tests measuringachievement at the end of college are available, yet somenormative studies of students year by year show a lowering(as introductory courses fade intothe background) rather than a raising of scoreswith continued time in college; one begins to question notonly the difficulty in contriving subject-matter tests fair to allstudents, teachers, and institutions, but also the validity of the academic way.There are those who call for the ultimate criterion ofsocial, personal, and professional achievement in life itself, and whohave found little or no relation- ship between academic successand success in life.The problem as to how the great massof the higher educated, going into a variety of work and other societalroles, can be evaluated is a most difficult one thatwill probably be resolved,eventually, by the separate formulation of avariety of measures for a varietyof purposes. Some ofthese measures may begenerally relevant to the role of the educated adult.(e.g., "social conscience"), and some may be peculiar to aparticular subgroup of thepopulation. Here 116

learn that there again, evenwithin occupationalfields, we shall we need avariety are fewultimate, universalqualities of goodness; of skills orsensitivities. of kinds ofdoctors with avariety of kinds

largely ignore thespecifi- One wayaround thesedilemmas is to successful comple- cation of avariety of impactcriteria, and to note into a neweducational tion of aneducational sequenceand movement indeed, or life sequenceas the truetest ofdevelopment. Perhaps, this is about after a half-centuryof testing andarticulation studies, that we all that we haveaccomplished. Thisis not a suggestion that we abandon the effort tospecify goals; butit is a suggestion of the impact may dobetter to leavesuch efforts tothe micro-view protest movement, of the particular course,instructor or peer, or four years at a par- rather than tothe macro-viewof what the total done to or for the ticular institutionof higher educationmay have individual.

implied in the Another way, whichis essentiallythe strategy forthcoming volume byNewcomb and Feldmn,n3 is to expand our knowledge of the variouseducational inputs,,process4s, and of context, and, asSteinn4 has foreseen, tofocus on the study traits. the interactions amongenvironmental forcesand student Personal life orcivilization itselfis a series ofprogressive hopefully, developments oradjustments of highercomplexity to, 117 some higher orderof functioning.It is my belief that at the very least we have reached a stage inhigher education where the researcher, rather than needing todemonstrate that members of the faculty at a particular institution do not recognize,and variously disagree with, the goal statement in thecatalogue, is and should be concerned with illuminating and specifyingthe many forces that make up an institution, and how these forces interactwith one another.

Directions for Future Research and Development Measurement specialists could retire comfortably tothe rose gardens behind their laboratories now if the implicationsof the history of the development of selection practices, and the supporting research, had given us an inventory of qualities important tosubsume in measurement devices, and a clean avenue for their acceptance and use. Some people are hidden away working on such devices; some of the ideas for the devices have comefrom intuition, some from analysis of the literature of one or another domain within psy- chology, and some from exhaustive factor analyses of existing tests. th4 But if we take the lessons of the past seriously, we must predict that little will come of the game of saying, "Here's a new concept that I know is important, "so let's measure it and quit fortoday."

There are, however, some areas of considerable promise for 118

the further developmentof the art and scienceof measurement, and for the educational systemas a whole.The first of these has to do with defining,elaborating, and improvingthe criteria by which For higher education,this has implications . students are evaluated. for the impact of thecollege on the student aswell as for the transition of studentsfrom high school tocollege. This problem goes beyondnoting what our presentpredictors in admissions studies tell us aboutthe nature of the criteria, orthe limitations of the criteria,altho`ugh this may be a point of entryfor the measure- ment /Specialistand a point of departurefor the new studies. My own belief isthat the problem calls for newapproaches and new partners. We are nowdealing not with the bestscholar's definition of a subject-matter field,but with his intuition, aswell as the intuitions of the significantsocietal leaders, as to thesocial and cultural utility of thatfield.

In our criterion construction, wemust go beyondassembling a core of expertsfor a three-day conference, orbeyond asking the faculty or societal leaders atlarge what the infinitelydesirable qualities of man may be.The job probably starts withthe best mealurement specialists andthe best teachers workingtogether with a variety of studentsin an on-going instructionalsituation, which focuses on aproduct-by-product evaluationand re-evaluation. Where the content of the newcriteria dimensions seemsflimsy, the 119 process thenbecomes that of confrontingmind and spirit with the search for more substantialqualities.It is also suspected thatin this work we shall not comeout with a list ofrelated qualities of goodness, but with avariety of frequentlyconflicting qualities. The latter, however, isprobably the better model formeeting the various role demands ofsociety.

A second area of promise,and one that is beginning toattract considerable attention,is that whichStein'15 called the "transactional" approach. He summarizesthis by stating: Basic to this approachis the assumption that success in college, as allbehavior, is a function of the trans- actions between theindividual and his environment.Indi- viduals affect and are affectedby their environments. Consequently, for purposes ofprediction it is important to understand both thecharacteristics of the individual and the environment.

We are experiencing amodicum of initial success withnew environmental assessmenttechniques as well as withtheoretical and operational studies ofstudent characteristcs otherthan intel- lectual ability.The work here is very much inthe elementary stages of development, but it appearsextremely promising.It not only subsumes convictions aboutselecting the student for the environment, but.also for modifying the environment sothat it may better serve the student. Some work with youngchildren has shown that one can use tests to form subgroups ofstudents who can be trainedby different 120 methods to the same criterion.If we can achieve this with cider students and with some of the more advanced kinds of studies, thenour educational institutions may support democracy in the more vital ways that democracy has always supported education.But whether weuse such work to determine the kinds of environments in which a group of bright students learn best, the different kinds of environmentsin which different subgroups of bright students learn best, or the set of environments that may permit some subgroups of students lower on the traditional ability hierarchy to learn as well as brighter subgroups inconventional systems--any of these appli- cations, could they be carr*ed off, would seem worthwhile.

The appeal of the work on cognitive styles (or "problem-solving styles") is attractive and promisinfor these kinds of reasons. One can argue that different disciplines require different kinds of solution strategies, and that effective educationis that which teaches the student new modes of attack.But here, again, the work must proceed hand-in-glove with the best subject-matter people the measurement specialist can muster to join him.

An important part of looking at promising avenueindicated by past research is to recognize that promise, defined by the measure- ment specialist, is not enough. Neither has educational practice always accepted the products of the measurement specialistonly for

triz 121

the most noble or pure of reasons.What chance for useful contribu- tion, indeed, does thisleave him? Certainly with more thanrushing to market a new test ofcreativity (one can be assured therewould be buyers, particularly among the naive,and that the naive would be particularly vulnerable to unquestioned acceptanceof the definition perpetrated by the test).

Measurement research is maturing tothe point where now it

may be moremission-oriented than discipline-oriented. Theachieve- ment, in our current academicworld, of a mission orientation is being brought about by the multi-teamapproach.McGra.c.h, had he lived two centuries ago, would have retired to amonkish cubicle and pontificated; today, his effort to defineinstitutional vitality is involving literally thousands of people and hundredsof perspectives, each carefully chosen and managed. The statisticalspecialist, although he may want to hide and develop a few dozen newmathe- matical models, is being pressed into practical serviceby the idea man with a problem; the test constructormust sit down with his client on the firing line. The experience ofworking together is

L beginning to show some promise of an ultimate commonlanguage and the prospect of real communication.

Far too many pages ago, we noted that in thebeginning was

Harvard.Is her light still shining in the darkness? It may be, but

' 122 it would seem that we arerapidly passing the stage, through measure- ment research activity, inwhich goodness is defined by an epitome institution (where the success ofgraduates may be assured by the success of their parents orby a universal stereotype of awe that may greet anyproduct).It would seem that we are moving toward an attempt to learn bystudies across all institutions as well asby focusing, where necessary, on discrete unitswithin an institution, as McKeachie'spaper116 in this series, for one, demonstrates forcefully.

Measurement research in higher education may havecontributed, most precisely, a relatively unitary dimension(scholastic aptitude) that has become an exclusive focus in admissions practices.This dimension was accepted because it saved time of faculty members, who otherwise would have devoted hours to test constructionand grading; and it has misled us into selecting into our systemsthose who can be taught with the least effort, involvement, ordifficulty. There is, to some readers of this report, some "proof" for such an interpretation.But others, perhaps of different biases, will see that what has been developed is a point of view that there is utility in attempts to specify some essential individual and institutional quality, and to test its meaning in some precise ways by studying its implications against other measures, and by looking, in inter- actional studies, for more than simple associations. The outcomes ,10

123 may produce some tests for educational consumers--but more than ever before, the pressures are for evaluating research not so much by its statistical niceties but by its (measured, of course !) impact on educational leaders and educational practice. Those who can accept a mission orientation honestly, who can learn to talk with and use those specialists from disciplines other than their own, who can use tools for their proper function--those people are opening their eyes to a magnificent dawning for the most exciting period of educational development yet.

'4`) .1.,10061

FOOTNOTES

1.The first part of this paper waspresented essentially in its present form at a conference onSelection Practices held at Grasmere, England, in April, 1967, co-sponsoredby the University of Lancaster zAnd the Institute of HigherEducation of Teachers College,Columbia University. 2.The major source for this sectionis a comprehensivehistorical account of admissionspractices until 1900 by E. C.Broome entitled "A Historical and Critical Discussionof College AdmissionsRequire- ments." Columbia UniyersityContributions to Philosullya_ay,21121,9.a. and Education, vol. 11, Nos.3-4. April, 1903.

3.Ibid., p. 204. 4.F. Bowles, "The Evolutionof Admission Requirements."C21.1eze Admissions: The Interaction of Schooland College, p. 24-36. rew York, College Entrance ExaminationBoard, 1956.

5.Ibid., p. 27.

6.Ibid., p. 25. 7. C. M. Fuess, TheColleaLBoard--Its First FiftYears. New York, Columbia University Press,1950. 8. College Entrance ExaminationBoard, Annual Re ort1965-66. 9. Fuess, 92. cit. 10. H. Chauncey, unpublishedaddress to Association of College Admissions Counselors, Highland Park,Ill. November 13, 1947. 11. W. S. Learned and B. D. Wood,The Student and His Knowledge, New York, The Carnegie Foundationfor the Advancement of Teaching, 1938. 12. College Entrance ExaminationBoard, Annual Re_port, 1945, p. 45. 125

13. Chauncey, ap.cit. 14. Learned andWood, 2.2. cit. International Stud 15.F. Bowles, Access toHi her Education: The of UniversityAdmissions. Vol. 1. NewYork, UNESCO and the International Associationof Universities, 1963.

16.H. Dyer, Review of"Admissions--College andUniversity," prepared for the forthcoming4th ed. of Encyclopediaof Educational Research. 17. D. Harris, "TheRelationship to College Gradesof Some Factors Other Than Intelligence."Archives of Psychology,vol. 20, No. 131. 1931." Factors AffectingCollege Grades: A Reviewof the Literature, 1930-1937." psychologicalBulletin, vol. 37, p. 125-66.1940.

18.j. A. Fishman and AnnK. Pasanella, "CollegeAdmission Selec- tion Studies." Reviewof Educa_tional Research, p.298-310. October, 1960.

19.J. A. Fishman, "SomeSocio-Psychological Theory forSelecting and Guiding College Students."In N. Sanford, ed.,The American Collcal and SocialInterpretation of Higher Learniaa, p. 666-89. NewYork, John Wiley, 1962. 20. M. I. Stein, PersonalityMeasures in Admissions.Research Monograph No. 5. New York,College Entrance ExaminationBoard, 1963.

21.D. E. Lavin, The Predictionof Academic Performance:A Theo- retical Analysis and Review ofthe Literature. New York,Russell Sage Foundation, 1965. 22. L. V. Koos, Private andPublic Secondar Education.Chicago, University of Chicago Press,1931. 23. Junius A. Davis and N.Frederiksen, "Public andPrivate School Graduates in College."journal of Teacher Education,vol. 6, p. 18-22.1955. 24. Audrey M. Shuey,"Academic Success of Public andPrivate School Students in Randolph-MaconWoman's College:I. The Freshman Year."Journal of Educational Research,vol. 49, p. 481-92. 1956. 41RT

126 25. C. C. McArthur, "Personalities ofPublic and Private School Boys." Harvard Educational Review,vol. 24, P. 256-62. 1954. "Subculture and Personality During theCollege Years."Journal of Educational Sociol_oaz, vol. 33, p.260-68.1960.

26.J. L. Holland, "The Prediction ofCollege Grades from the California Psychological Inventoryand the Scholastic Aptitude Test." ournal of Educational Ps cholo vol. 50, p. 135-42. 1959."The Prediction of College Grades from Personalityand Aptitude Variables." Lurnal of Educational Psycholocia, vol. 51, p.245-54. 1960.

27.J. W. Getzels, "Non-IQ Intellectual andOther Factors in College Admission."In K. E. Anderson, ed., TheCornir in the Selection of Stud2atsforCollege Entrance, p.23-28. Wash- ington, American Educational ResearchAssociation, 1960. 28. Junius A. Davis, "What College TeachersValue in Students." Colle e Board Review, No. 56, p. 15-18.1965. 29. Getzels, op_. cit., p. 28. 30. W. Coleman and E. E. Cureton, "Intelligenceand Achievement: The 'jangle Fallacy' Again." Educational andPsychological Measure- ment, vol. 14, p. 347-51. 1954.

31.E. B. Page, "The Imminence of Grading Essays byComputer." Phi Delta Kappan, vol. 47, p. 238-43. 1966. 32.S. Klein and R. Skager, "Sontaneirateness." as a Dimension of Esthetic Judgment,Research Bulletin 66-14. Princeton, N. J., Educational Testing Service, 1966. 33. Junius A. Davis, "Non-intellectual Factors in College Student Achievement." In From High School to CollecuLi- selorsp. 72-81, esp. p. 73. New York,College Entrance Examina- tion Board, 1965.

34.D. R. Saunders, Moderator Variables in Prediction withSpecial Reference to Freshman Engineering Grades and the Strong VIB. Research Bulletin 53-23.Princeton, N. j, Educational Testing Service, 1953. 35. N. Frederiksen and S. D. Melville, "Differential Predictability in the Use of Test Scores."Educational and Ps cholo ical Measure- ment, vol. 14, p. 647-56. 1954. 127

36.E. E. Ghiselli,"Differentiation of Individuals inTerms of Their Predictability." Journal ofApplied Psychology, vol. 40, p.374-77. 1956. 37.J. W. French,Manual for the Exerimentarative Prediction Batteries.Princeton, N. J., EducationalTesting Services 1964. 38. G. E. Sch lesserand J. A. Finger,"Non-intellective Predictors of Academic Success inSchool and College."School Review, vol. 73, p. 14-29.1965.

39.S. Messick, "PersonalityMeasurement and CollegePerformance." In Proceedings ofthe 1963 InvitationalConference on TestingProblems, p. 110-29.Princeton, N. J., EducationalTesting Service, 1963. 40. Fishman, 22: cit.

41.Messick, oa. cit. 42.Dyer, 22 cit. 43. Bowles, Access toHigher Education, op... cit. 44. Dyer, ga. cit. 45.See, for example, J. G.Darley, Promise and Performance:A Studof Abilit and Achievement inAmerican Hi her Education. Berkeley, Center for the Study ofHigher Education, Universityof California, 1962. A. W. Astin,"Distribution of Students Among Higher Educational Institutions."ournal of Educationaj._ayol?o1_ogz, vol. 55, p. 276-87. 1964. Astin,Who Goes Where togaufaE2Chicago, Science Research Associates,1965. R. C. Nichols, "College Preferences of Eleventh GradeStudents." NMSC Research Reports, vol. 2, No. 9.1966. A. W. Astin, R. J. Panos,and J. A. Creager, "National Norms for EnteringCollege Freshmen--Fall, 1966.ACE Research Reports, vol. 2, No. 1.1967. College Entrance Examina- tion Board, Manual of FreshmanClass Profiles1967-69. New York, CEEB, 1967. 46. Junius A. Davis, "The CriterionProblem in College Admissions Research." In j. M. Dugg6.1,ed., Research in Hi herEducation: Guide to Institutional Decisions, p.25-34. New York, College Entrance Examination Board,1965. 47.K. E. Eble, The ProfaneComedy. New York, Macmillan,1962. Hanford, "Testing:Wise Restraints."In L. Wilson, 48. G. H. 225-27. ed., EmergingPatterns in AmericanHi her Education, p. Washington, AmericanCouncil on Education,1965. 49. Davis,"Non-intellectual Factorsin College StudentAchieve- ment," 22: cit. ,p. 79. College 50.J. W. Trent, "ANew Look atRecruitment Policies." Board Review,No. 58, p. 7-11.Winter 1965-66. 51. Learned andWood, 222. cit. on Pre- 52. Junius A.Davis, Distributionof 1957 Entering...Freshmen Admissions Indices.Atlanta, Board ofRegents of the University System of Georgia,1958. of this work, see D.G. Paterson and J.G. Darley, 53. For a summary Press, Men, Women,ancilobs. Minneapolis,University of Minnesota 1936. Change. New York, 54.T. M. Newcomb,Personality and Social Dryden, 1943. 55. C. R. Pace,They Went to College.Minneapolis, University of Minnesota Press,1941. 56. D. Chamberlin,E. Chamberlin,N. E. Drought, and W.E. Scott, Did TheySucceed in College?New York, Harper,1942. Explora- 57.P. L. Dressel andL. B. Mayhew,General Education: tions in Evaluation.Washington, AmericanCouncil on Education, 1954.

58.B. S. Bloom, ed.,8._a_ikcorn.z..o. jEducationalOb'ectives: pa..L.t_Titl-itive Domain.New York, DavidMcKay, 19. description of thisinstrument, see S. R.Hathaway, 59. For a general Kaplin, "The MinnesotaMultiphasic PersonalityInventory." In 0. J. ed., En.g.zalopecliaof VocationalGuidance. New York,Philosophical Library, 1948.

60.J. P. Guilford and W.S. Zimmerman,Manual for the Guilford- Zimmerman TemperamentSurvey. BeverlyHills, Calif. , Sheridan Supply Company, 1949 129

Exploratory 61.P. E. Jacob, ghar_ CollezeL of the Im act ofColle e Teachin.New York, Harper,1957.

62.Sanford, op_:, cit. 63. A. W. Astin,The College Environment.Washington, American Council on Education,1968.

64.T. M. Newcomb andK. A. Feldman,Thelynpactsof' Colleges upon TheirStudents. Forthcoming. 65. M. Keetonet al., a forthcomingvolume probably to carrythe title of the study, TheFuture of Liberal ArtsColle es.

66.B. R. Clark and M.Trow, "The OrganizationalContext." In T. M. Newcomb andE. K. Wilson, eds.,Collrzips.; Problems and ,Pr:onpects forResearch, p. 17-70.Chicago, Aldine, 1966.

67.Sanford, op_:. cit. 68. H. S. Dyer, "Caninstitutional Research Lead to aScience of Institutions?" EducationalRecord, vol. 47, No.4, p. 452-66. Fall, 1966. 69. W. H. Cowley, "Twoand a Half Centuriesof Institutional Research." In R. G. Axtand H. T. Sprague,eds., College Self Interstate Commission It24cal,p. 17-22.Boulder, Colo., Western for Higher Education.,1960. 70. Harris, oat cit. 71. Fishman andPasaneila, ap... cit.

72.J. Re Hills, "AdmissionsProcedures that Make Sense."In Duggan, ap..! cit., p.16-24. 73. Learned and Wood, oatcit.

74.Darley, 224 cit. 75. Astin, Who GoesWhers.e.ge, c2p..z cit. 76. R. L. Thorndike andElizabeth Hagen, 10,000Careers.New York, John Wiley, 1959. .401,

130 77. D. E. Super, "A Theory of Vocational Development." American psychohaist, vol. 8, p. 185-90. May, 1953. 78. Naomi Stewart, "AGCT Scores of Army Personnel Grouped by Occupation." Occupations, vol. 26, p. 5-41.1947. 79. James A. Davis, "Reference Group Processes and the Choice of Careers in Science." Paper presented at 1964 convention of the American Psychological Association in Los Angeles.

80.Ibid.

81.S. C. Webb, "Measured Changes in College Grading Standards." Collge Board Review, No. 39, p. 27-30.1959.

82. Louis R. Aiken, Jr., "The Grading Behavior of a College Faculty.11 Educational and tuchologicallyeasurement, vol. 23, No. 2, p. 319-22. 1963. 83.Davis, "Reference Group Processes and the Choice of Careers in Science." p_at cit. 84.B. G. Fricke, Opinion, Attitude and Interest Survey Handbook. Ann Arbor, University of Michigan, Evaluation and Examinations Division, 1963. 85. Paul A. Heist et al., Omnibus PersonalitInventory_Research Manual. Berkeley, Center for the Study of Higher Education, Uni- versity of California, 1962. 86. Manual for the Edwards Personal Preference Schedule. New York, The Psychological Corporation, 1959. 87. G. G. Stern, Scoring Instructions and College Norms for the Activities Index and the College Characteristics Index. Syracuse, N. Y., Syracuse University, Psychological Research Center, 1963. 88. H. A. Murray, Explorations in Personality. New York, Oxford University Press, 1938.

89.Stern, opz cit.

90. Schlesser and Finger, 224 cit. 131

91.For a summary of thesestudies, see "Tenth Annual Review of Research," NMSC ResearchReports (vol. 2,No. 11, 1966), published by the National MeritScholarship Corporation. 92. E. D. Farwellet al."Student Personality Characteristics Associated with Groups ofColleges andFields of Study." gontat and University, vol. 37, p.229-41. 1962.P, A. Heist et al. "Personality and Scholarship."In W.W. Charters, Jr., and N.L. Gage, eds., Readings inthe SocialPsychology of Education, p. 65-73. Boston,Allyn & Bacon,1963.

93.F. Dameron and S. Messick,Response Styles and Personality Variables: A Theoretical Integrationof Multivariate Research. Research Bulletin 65-10.Prin eton, N. J., EducationalTesting Service; 1965. 94. Clark and Trow, 22... cit. 95. R. E. Peterson, Technical ManiclentQuestion- naires.Princeton, N.J., Educational Testing Service,1965. 96. R. E. PetersonSome Bio rahical and Attitudinal Character- istics of Entering Colle e Freshmen. ResearchBulletin 64-63. Princeton, N. J.,Educational Testing Service, 1964. 97. R. I. Panos, A. W. Astin,and J. A. Creager,"National Norms for Entering College Freshmen--Fall, 1967."ACE Research Reports, vol. 2, No. 1967. 98. College StudentProfiles. Iowa City, Iowa, AmericanCollege Testing Program, 1966. 99. DBoroff, garnasf_ITSA. NewYork, Harper Bros. , 1961.

100.Anne Roe, "A Psychological Studyof Eminent Physical Scientists." Genetic clyIc.ozynigonoqraphs, vol. 43, p. 121-239.1951.

101. A. W. Astin, "An EmpiricalCharacterization of Higher Educa- tional Institutions." jurna1 ofEducational Psychology, vol. 53, p. 224-35. 1962.

102.J. M. Richards, Lorraine Rand,and L. P. Rand, A Descriztaki 21Junior Colleges. ACT ResearchReport No. 5. Iowa City, Iowa, American College Testing Program,1965. 132

103. C. R. Paceand G. G. Stern"An Approach to theMeasurement of PsychologicalCharacteristics of CollegeEnvironments." Journal of EducationalPsychology, vol. 49, p.269-77. 1958.

104.Murray, 22t cit. 105. G. G. Stern,"Environments for Learning."In Sanford, oat cit., p. 690-730. 106. C. R. Pace Colle e and UniversitEnvironmental Scales. Princeton, N. j.,Educational Testing Service,1963. 107. A. W. Astin andJ. L. Holland, "TheEnvironmental Assessment Technique: A Way toMeasure CollegeEnvironments." J2rnalof Educational Psychology.,vol. 52, p. 306-16.1961. 108. A. W. Astin,"Further Validation ofthe EnvironmentalAssess- ment Technique." oulL_:L.Ial of Ed.....1_,Icatior..2113.sycho,vol. 54, p. 217-26.1963. 109. Astin, The Colle eEnvironment, 224 cit.

110.R. E. Peterson and D.E. Loye, eds.,Conversations Toward a Definitionof Institutional Vitality,.Princeton, N. J., Educational Testing Service, 1967. and the Innovative 111. J. W. Gardner,Self-Renewal: The Individual Society. New York,Harper & Row, 1964.

112.Newcomb and Feldman, 2.24cit.

113 Ibid.

114.Stein, 22z cit.

115.Ibid. 116. W. J. McKeachie,"New Developments inTeaching." In E. H. Hopkins, ed.,New Dimensions inHigher Education. No.16. Washington, U.S. Officeof Education, 1967. ANNOTATED BIBLIOGRAPHY

A. The Phiiosophy and Practice of SelectiveAdmissions

1.Bowles, F., Access to Higher Education:The International Study of University Admissions. UNESCO andthe International Association of Universities, vol. I. New York,1963. Drawing on his fifteen years' experience aspresident of the College Board, as well as upon a two-yearstudy of university admis- sions in a number of countries throughoutthe world, this scholarly study relates education and its functioning tothe public purpose, and treats the college admissions process as abroad, social phe- nomenon, responsive todeep-seated national pressures and aspira- tions within a country.

2. ,The Refounding of the College Board,1948-1963. New York, College Entrance ExaminationBoard, 1967. "An informal commentary and selected papers"documenting the activities of the College Entrance ExaminationBoard under Bowles's period of directorship.This period is particularly crucial, for its beginning marked the establishment of EducationalTesting Service as the technical arm of theCollege Board. The collection chronicles the development of modern perceptions ofselection needs and policies in the period of rapid development andapplication of admissions testing. 3.Broome, E. C., A Historical and CriticalDiscussion of Colle e Admissions Re uirements .Columbia University Contributions to Philosophy, Psychology, and Education, vol. XI, Nos.3-4. April, 1903.(Reprinted 1963, College Entrance ExaminationBoard.) A -.A.assic and scholarly (in the oldsense) study of the evolution of admissions practices as a function of theevolution of institutions of secondary and higher education. A "must" for theeducational historian as well as for the college admissions officer.

h "LI. ;%., "..-IrqTmegr-

134 4. College Entrance Examination Board, College Admissions. Vols. 1-10. One volume per year from 1954 to 1963; published in New York Joy the College Entrance Examination Board. For a ten-year period the College Entrance Examination Board held an annual invitational. conference for a small group of admissions officials on selected problems in college admissions. Speakers were chosen with special care; this set of volumes presents their papers. 5.Duggan, J. M., and P. H. Hazlett, Predicting College Grades. New York, College Entrance Examination Board, 1961. A workbook for the non-initiated that provides cookbook formulas and worksheets for handling a prediction problem.

6.Dyer, H. S., "Admissions--College and University."To appear in R. Ebel and Victor Noll, eds., Encyclopedia ofEducational Re- search. (4th Ed.)Forthcoming from the American Educational Re- search Association. A perceptive review of problems, practices, and research in admission of students to college, this sweeping review isas appro- priate for the layman as for the sophisticated researcher. 7.ruess, C. M., The Colle e Board--Its First Fifty Years. New York, Columbia University Press, 1950. A folksy account by a former private school headmaster active for many years with the College Board, this volume describes (generally accurately, always delightfully) the first fiftyyears of development of the College Entrance Examination Board. 8.Thrasher, B. A., Colle e Admissions and the Public Interest. New York, College Entrance Examination Board, 1966. This book contains a series of literate reflections on the admis- sions process by B. Alden Thrasher, for twenty-five years director of admissions at M.I.T. Thrasher is particularly concerned with the social forces that push students into higher education. Theo- retical and philosophical analyses are quite keen; the volume is less satisfying in its discussion of how these issues may be handled in practice. ,

135 B. Reviews of Research on the Prediction of Success in College

9.Davis, Junius A., "Non-intellectual Factors in College Student Achievement." In From Hi2h School to College: Readings for,poun- selors, p. 72-81. New York, College Entrance Exak..nation 1965. Although directed to pre-college counselors in the secondary schools, this general review concludes that there is not much beyond conventional aptitude and academic achievement for the high school counselor to use.

10.Fishman, J. A., and Ann K. Pasanella, "College Admission Selection Studies." Review of Educational Research, p. 298-310. October, 1960. This paper, directed toward the moderately technical reader, reviews 580 studies during the decade from 1949 to 1959.Its useful bibliography contains fifty-seven references.

11.Garrett, H. F., "A Review and Interpretation of Investigations of FactDrs Related to Scholastic Success in Colleges of Arts and Science and Teachers Colleges." journal of Experimental Education, vol. 18, p. 91-138. 1949. This report is a review of prediction studies over a twenty-yerlr period beginning in 1930, for which some 194 studies are mentioned.

12.Harris, D., "Factors Affecting College Grades: A Review of the Literature, 1930-1937." psychological Bulletin, vol. 37, p. 125-66. 1940.

13. ,"The Relationship to College Grades of Some Factors Other Than Intelligence." ..23-ays...§ol_Ps ci_y_12Loayr, vol. 30, No. 131.1931. Both articles by Harris are reviews of prediction studies published over the period indicated. He focuses on the attempts to find corre- lates of ;.chievement in college beyond the conventional cognitive measures. His summary indicates that those who do this kind of study are generally a rather haphazard lot.

14.Lannholm, G. V., "Review of Studies Employing GRE Scores in Predicting Success in Graduate Study, 1952-1967." Graduate Record Examinations S ecial Re ortNo. 68-1.Princeton, N. J., Educational Testing Service, 1968. 77."4

13 6

A review of some thirty-six published and unpublished studies which employed GRE scores in predicting success in graduate study.

15.Lavin, D. E., The. Prediction of Academic Performance: A Theo- retical Anal sis and Review of the Literature. New York, Russell Sage Founda.tion, 1965. The most recent general review of the literature on the prediction of student performance, this volume is noteworthy for its grasp of the broader issues as well as for its treatment of sociological and socia1-psycho1oT:;a1 factors that affect levels of achievement.

16.Stein, M. I., Personality Measures in Admission. New York, College Entrance Examination Board, 1963. This report is an excellent summary of a review commissioned by the College Board of the use of personality measures in college admissions.Its major contribution lies in its analysis of curt-Ant failures and in implications for future studies. This volume contains a useful bibliography.

C. The Evaluationi Through Measurement Perspectives, of Higher Education

17.Brumbaugh, A. J., Research Desined to Imrove Institutions of Higher Leaning.. Washington, American Council on Education, 1960. Although measurement is not a matter cf particular concern in this little guide, a useful handbook for the general administrator con- cerned with institutional self-studies is provided.

18.Dressel, P. L., et al., Evaluation in High,er Education.Boston, Houghton Mifflin, 1961. This volume is a collection of essays by Dressel and his asso- ciates at Michigan State University on applications of measurement and evaluation procedures and points of view to the administration of the institution and its programs.It provides an excellent overview, as for a beginning graduate student in a formal study of higher educa- tion, of problems ranging from selection and placement of students, through evaluation of growth of students in various areas, to evaluation of instruction or institutional self-study.

M14 137

19.Dressel, P. L., and L. B. Mayhew, General Education:Explora- tions in Evaluation. Washington, American Council onEducation, 1954. This volume sammarl.zes for the intelligent layman the results of a four-year cooperative study of the impact of the general college programs in nineteen colleges and universities.Concluding chapters on implications and unresolved issues,and suggestions for future studies, are particularly noteworthy. 20. Lazarsfeld, P. F., and S. D. Sieber, Organizing Educational Research. Englewood Cliffs, N. J., Prentice-Hall, 1964. Although drawn from the perspective of sociology rather than measurement pse, and frequently critical of measurement scientists (especially those outside the colleges), this volume is a useful and thoughtful summary of the problems in modern educational research.

D. Current Standard Texts in Tests and Measurement 21. Chauncey, Henry, and John E. Dobbin, Testing: Its Place in Education Today,. New York, Harper & Row, 1963. Directed toward parents and teachers, this informative little volume cuts through many of the sources of popular fallacies and confusion about the place of testing in education today. A central notion has to do with the test as a "partner of teaching." Although directed toward the public school setting, many sections are quite useful for students of higher education. 22. Cronbach, Lee j., Essentials of Psychological Testing.2nd ed. New York, HarperBrothers, 1960. This book is far and away the most popular undergraduate text in general psychological testing. 23. Jackson, Douglas N., and Samuel Messick, eds., Problems in Assessment. New York, McGraw-Hill, 1967. A monumental collection, through seventy-four chapters and almost a thousand pages, of classic studies as well as modern state- ments of contemporary issues.It is directed most precisely at the c. graduate student or measurement research specialist.

_ rtr

138

24. Lindquist, E. F.,ed., Educational Measurement. Washington, American Council on Education, 1951. Although now out-of-date (a revised editionis being prepared under the editorship of R. L.Thorndike), this volume is the classic reference for the planning, construction, use,and analysis of the educational test. 25. Linn, R. L., J. A. Davis, andPatricia Cross,A Guide to Research Design. Princeton, N. J.,Educational Testing Service, 19 65 Focusing principally on tests available fromETS for general institutional research purposes, this manual iswritten for the insti- tutional researcher who has entered that rolefrom some background other than social science research or statistics.It includes appen- dices on statistical terms and procedures,and provides more than one hundred references. 26. Nunnally, jum C., Jr., Tests andMeasurements: Assessment and Prediction. New York, McGraw-Hill,1959. This volume, like the Cronbach volume, is apopular undergraduate text in test and measurement.It is particularly useful in terms of its concise and lucid treatment of statisticalproblems in the use of testing. 27.Stuit, D. B., G. C. Helmstadter, and N.Frederiksen, Survey of Colle e Evaluation Methods and Needs:4 Re ort to the Came ie gszooration. Princeton, N. J., Educational TestingService, 1956. This out-of-print report is a summary of methodsand materials for evaluating, in self-studies, the following aspectsof a college: institutional objectives, curriculum, faculty,instructional effective- ness, student body, andstudent personnel services. After attempts to define the evaluation problem in termsof underlying dimensions in each area, the report cites both availableand needed methods and materials.

E. Suszasof Student Input Dimensionsof Diversity. Amon Institutions of Higher Education 28. Astin, A. W., R. J. Panos, and J. A.Creager, " National Norms for Entering College Freshmen--Fall,1966." ACE Research Reports, vol. 2No. 1.1967. Washington, American Council onEducation, 1967.

:;tr 139

Drawing from Astin's new concern with student behavior as an indication of the learning climate of an institution, thisreference source provides a good answer tothe question, "What are students like today?" The data come from a survey of almost 300,000entering freshmen students at a carefully selected sample of 359colleges and universities in 1967.Distinctions among various types of institutions are presented. 29. College Entrance Examination Board, Manual of FreshmanClass Profiles1967-69. New York, College Entrance Examination Board, 1967. Intended as a source book for secondary school counselors and others who help students make their college plans, the statistics on tested ability and high school performance of entering freshmen classes have served a iiumber of research studies well. This sixth edition contains profiles supplied by 520 member colleges of the College Board. 30. College Entrance Examination Board, Manual of Freshman Class Profiles for Indiana Colleges. New York, College Entrance Examination Board, 1965. This volume is significant because it attempts to demonstrate the feasibility of augmenting the College Board's National Manual of Freshman Class Profiles by including, in addition to test data and a self-description for each college, the results of formal environ- mental assessment studies and data on the kinds of students entering the college.It is published as a guide for high school counselors and students.

31.Dar ley, J. G., Promise and Performance: A Study of Ability...ad Achievement in American Higher Education. Berkeley, Center for the Study of Higher Education, University of California, 1962. Dar ley's inquiry is concerned with the present structure of higher education in the United States in terms of ability, performance, attrition, and occupational plans of students in a national sampleof institutions. Contrasts between freshmen in 1952 and 1959 are pre- sented, with intensive analyses of students in Minnesota, Wisconsin, Ohio, and Texas. 32.Hills, J. R., Counselor's Guide to Geor ia Calle es. Atlanta, Office of Testing and Guidance, Board of Regents, University System of Georgia, 1965. 140

An updating of anearlier survey by J. A.Davis, this manual presents admissionsdata on enteringcollege freshmen,together with procedures forprediction of grades, inthe public and private colleges of Georgia. 33. Learned, W.S., and B. D. Wood,The Student and HisKnowleaI e: A Report to theCarnegie Foundation onthe Results of theHigh School and Colle e Examinationsof 19281930 and 1932. NewYork, Carnegie Fund for the Advancementof Education, 1938. The classic study ofeducational developmentwhich examined achievement levels oversecondary schools andcolleges in the state of Pennsylvania bythe use of educationaltests, this investigation provided the first positiveindication of the extentof the diversity that exists both amonginstitutions and amongdepartments -Alf.ithin institutions. 34. Peterson, R. E., Manual, e stu ire s .Princeton, N. J.,Educational Testing Service,1965. Intended as a guide for usersof the College StudentQuestion- naires, this manualprovides normativeinformation on the rangeof background factors, aspirations,and experiences ofcollege students. 35. Seibel, DeanW., A Studof the Academic Abilitand Performance aLlunior Colleclents.Princeton, N. J., EAS FieldStudies Report, Educational Testing Service,1965. This study is a follow-upof a represt.ntativenational sample of high school seniors forwhom ability measures(from the Preliminary Scholastic AptitudeTest) were available. Data arepresented which describe the academicability of students whoenroll in two-year institutions and studentswho enroll in four-yearinstitutions accord- ing to their performanceduring the first year ofcollege.

The Anal sis of theLearnin Context orthe Colle e Environment 36. American CollegeTesting Program,ColleaSdent Profiles. Iowa City, Iowa,American College TestingProgram, 1966. This volume, preparedby the ACT Research andDevelopment Division, is an extensivedescription of studentsenrolled in colleges and universities using theACT program.Statistical data include information on a wide range ofstudent characteristics.A testimony 141 to the diversity that exists amongAmerican institutions of higher education, this volume is of interest to anyonewith patience to examine its data, and who isinterested in contrasting his institution with others or is concerned withthe broad range. 37. Astin, A. W., The ColleEnvironment.Washington, American Council on Education, 1968. This report, heavily based on data collectedthrough the National Merit Scholarship Corporation, is anexcellent analysis of the problems and prospects in environmental assessment.Its implications are directed both toward the general collegeadministrator and towdrd the individual psychologist or teacher concernedwith the impact of the environment on human development.

38. ,"Distribution of Students Among HigherEduca- tional Institutions." Iournal of EducationalPsychology, vol. 55, p. 276-87. 1964.

39. ,Who Go3s Where to Colle e? Chicago,Science Research Associates, 1965. Astin has provided a useful guide forpre-college counselors as well as a set of environmental measuresthat are attracting wide use in cross-institutional studies. Usingdimensions developed by the National Merit Scholarship Corporation team,Astin provides profiles for more than a thousand institutions.His dimensions.are drawn from statistical combinations of facts of recordabout the institutions. I! 40.Barton, A. H., Oroni.zational Measurementand its Bearin on -the Study of College Environme=. New York,College Entrance Examination Board, 1961. This review is a sweeping andlandmark summary, commissioned by the College Board, of sociologists' experiencein environmental assessment, with implications for carryingthe problems and procedures to assessment of the college environment.

41.Pace, C. Robert, Analyses of a National Sampleof College Environments. Washington, U.S. Office ofEducation, 1967. This volume describes work done by Pace in aUSOE study of 175 colleges and universities in 1964 and 1965.Some useful contrasts between the author's approach and that of Astin areincluded. :^!..

.

142

42. ,Colle e and Universit Environment Scales: Technical Manual. Princeton, N. J., Educational Testing Service, 1963. This manual, prepared as a guide for users of Pace's instrument, contains the rationale and much descriptive data on the American college environment. A revised edition, now in manuscript form, will shortly be available; that edition will augment the original data by incorporating information from many of the institutions studied in the NORC "glow-up of college graduates. 43.Peterson, Richard E., The Scope of Or anized Student Protests in 1964-1966.Princeton, N. J., Educational Testing Service, 1966. This volume reports the results of a study of institutional factors which are popularly assumed to have implications for incidents of student protest. A number of myths are exploded and a number of hidden factors are revealed. The study involved the majority of four- year colleges and universities across the country. 44.Stern, G. G., "Characteristics of the Intellectual Climate in College Environments." Harvard Educational Review, vol. 3, p. 5-41., 1963. This paper probably represents the best review available today of the author's approach in measuring the college environment. Although semi-technical, it io presented in a form that the intelligent layman may understand.

G. The Im of Collecies U on their Students 14.7.4..11 45. Davis, James A., Great Aspirations. Chicago, Aldine, 1964.

46. ,Under raduate Career Decisions: Correlates of Occupational Choice.Chicago, Aldine, 1965. These two volumes are the first and second reports of a sweeping follow-up of graduates from some three hundred American colleges and universities. The data are of interest not only in their own right, but also because other investigators may obtain them from NORC after that organization has completed its ii;;Itial analyses. As a data bank for a continuing inquiry, this is a highly significant effort.

47. Newcomb, T. ,Personality and Social Change. New York, Dryden, 1943. 143

This classic study found that youngladies from conservative Republican backgrounds moved in theirviews toward those of their faculty in a college where mostinstructors were liberal Democrats. It is significant for the solidevidence it presents on the more subtle issue of attitude change ofcollege students as a func;tion oftheir experiences in college.

48. ,and K. A. Feldman, The Im acts ofColle es pon Their Students.Forthcoming. This volume, now in pre-publicationdraft, is the result of a Carnegie-supported exhaustive review of avariety of research projects, both published and ongoing, that haveimplications for defining, measuring, and manipulating the impactof colleges upon their students. It may well become, when published,the most significant contribution L_higher education research of this decade. 49. Pace, C. R., :they.1/e.Minneapolis, University of Minnesota Press, 1941. A follow-up of former Universttyof Minnesota General College students, this study was an earlyclassic toward Toviding information back to teachers and administratorsthat might be used to modify the higher education experience.

H. Research Reports fromOrganizations Concerned with Measurement Studies ofHiclaerklucation 50. American College TestingProgram, ACT Research Re orts. This series was begun in 1965 by staffof the Research and Development Division of the American CollegeTesting Program in Iowa City, Iowa. Reports are issued asstudies are completed. Generally they involve analyses of datacollected in the ongoing programs of that organizationtoward particular problems, such as educational goals of entering freshmen, orthe relationship between academic and non-academic accomplishment. 51. American Council on Education,ACE Research Reports. Published from time to time by the Office ofResearch, American Council on Education in Washington,these reports draw principally on data from a major,ongoing, and multi-purpose study of a sample of colleges and universities.

c0t 144

52. College EntranceExamination Board, College BoardReview. This "slick" journal ispublished quarterly; its principaltargets are.college admissions officersand high school guidancecounselors. Papers are literate andnon-technical; important CEEBresearch projects as well asthink pieces about the admissions orguidance process are included. 53. Educational TestingService, Research Bulletins. Viewed internally as apre-publication issue, thesebulletins report the results of majorstudies done by the ETSresearch staff as those studies are completed.The RB series is generallydirected at the sophisticated researcherwithin the area of the subjectof the study. A listing of researchbulletins of general interest iscontained in the ETS Annual Re ort. 54. National Merit ScholarshipCorporation, NMSC Research Re orts. This excellent series wasinitiated in 1965 by R. C.Nichols. With a prospective phasing outof research activities atNMSC it is now threatenedwith extinction. Reports aredirected toward the professional audience. Seeespecially krEsQ Research Re ort1966, vol. 2, No. 11, "Tenth AnnualReview of Research," whichcontains not only work in progressand completed that year, butalso provides abstracts of studies completed atNMSC from its founding in1955. In all, 130 papers or projects arelisted.

Major Professional journalsContainina Reports of Measurement Studies American Educational ResearchTournal American ournal of SociologI American Sociological Review British ournal of EducationalPsychology College and University

c212-s_un

EAS Resources (ETS) V '4+1'.

14'

Education Reca s (ETS) Educational and Ps cholo.ical Measurement

Educational Record Harvard Educational Review lournal of Applied Psychology ournal of College Student Personnel

Journal c.A_Cour_zelka_i P_sachc_?Lnay. Journal of Educational Measurement Journal of Educational Psychology Journal of Educational Sociology Journal of Psychological Studies Personnel and Guidance Journal Psychological Abstracts

cl,z_calogical Bulletin Psychological_zalogapla

Dapils_p_ata Re ister of Research Pro ects in H' her Education (Society for Research into Higher Education, Ltd., 2 Woburn Square, London, W.C.l) The Research Reporter (Center for Research and Development in Higher Education, University of California at Berkeley) Review of Educational Research School and Society Science

Sociology of Education V0,

14 6 Organizations or Research Centerswith Measurement Research Teams inHiher Education

American College Testing Program (Iowa City, Iowa) American Council on Education (Washington, D. C.)

Bureau of Applied Social Research (Columbia University) Center for Research and Development inHigher Education (University of California at Berkeley) Center for Research on Learning and Teaching (University of Michigan) Center for the Study of Evaluation of InstructionalPrograms (University of California at Los Angeles) Center for the Study of Higher Education (University of Michigan) Centre for the Study of Higher Education (University of Lancaster, England) College Research Center (Vassar College) Educational Testing Service (Princeton, N. j.) institute of Education (University of London, England) Institute of Higher Education (Teachers College, Columbia University) Institute for Social Research (University of Michigan) National Merit Scholarship Corporation (Evanston, Ill.) National Opinion Research Center (University,.of Chicago) REACTIONS

In order for this second series of "New Dimensions inHigher Education" to better serve the needs of colleges anduniversities throughout the nation, reader reaction is herewith being sought. In this instance, with respect to ..Uplications of the Scienceof Measurement to Hi her Education, the following questions are asked: 1. Can you suggest other completed research, the results of which would add significantly to this report? 2. What problems related to this subject should be given the highest priority, in terms of further research? 3. What can the United States Office of Education do to encourage and support constructive innovation and change, based upon recent developments in the science of measurement?

Kindly address reactions to:

Dr. Winslow R. Hatch Bureau of Higher Education Research Office of Education U. S. Department of Health, Education, and Welfare Washington, D. C.20202

egf