WIXLIB

Chapier. 111Figure 4.1 Attendance of twelfth-grade students at SAT sessions, 1956-57through 1968-69Figure 4.2 Ratio of number of SATS taken by twelfth-grade students tonumber of births 18 years earlier, 1956-57 through 1968-69.Figure 4.3 Attendance of twelfth-grade students at Achievement Tests.ssions, 1956-57 through 1968-69Figure 4.4 Percentage of Achievement. Tests administered in five subjectmatter categories, 1956-57 through 1968-69Table 4.1Table 4.2Table 4.3Table 4.4Table 4.5Scholastic Aptitude Test means and standard deviations for aAim:al sample of secondary school seniors tested in 1960,ry sec and academic status in 1960-61sAT- verbal and SAT-mathematical mean scores for applicants,accepted applicants, and enrolled students reported by 18collet -s in the Manual of Freshman Class Profiles of 1962.1964. and 196567Numbers cf colleges in which changes of various amountsoccurred in percentile ranks for selected scaled scores b'tweenthe 1963 edition and the 1965-67 edition of the Manual ofFreshman Class ProfilesMeans, standard deviations, and numbns of ScholasticAptitude Test scores, classified by candidate's sex andeducational level, 1967-68 and 1968-69Means, standard deviations, and numbers of cases (inthousands) for all candidates taking the SAT, by year,1956-57 through 1965 -69909193111838")888992Table 4.6 Percentages of Achievement Test candidates taking the varioustests at !east once during the last two years of secondary school 94Table 4.7 Scholastic Aptitude 'rest means and standard deviations for95Achievement Test candidatesTable 4.8 Percentages of Achievement. Test candidates tested in thejunior year, senior year, or both years, by specific96Achievement TestTable 4.9 Means, standard deviations, and numbers of AchievementTest scores, classified by to t and by candidate's sex ant'educational level, 1967-68 and 1968-6997Table 4.10 Mean differences in Achievement Teot scores earned by boys100and girls. 1%7-68 and 1963-69Table 4.11 Mean differences in Achievement Test scores earned by juniors101and seniors, 1967.68 and 1968.69Table 4.12 Means, standard deviations, and numbers of scorn- onAchievement Tests in foreign languages, classified by test andby candidate's sex, educational level, and number of semestersof study of the language in secondaiy school, 1967.68 and1968-691029

Chapter IV ((70N7 'NUMTable 4.13 Mean differences in Achievement Test scores in foreignlanguages earned by candidates with different amounts of107language study, 1967-68 and 1968-69Table 4.14 Means, standard deviations,, and numbers of scores onAchievement, Tests in mathematics, classified by test. and bycandidate's sex, educational level, and number of semesters ofstudy of secondary school mathematics, 1967-68 and 1968-69108Table 4.15 Mean differences in Mathematics Level I Achievement Testscores earned by candidates wtita different amounts of110mathematics study, 1967-68 and 1968-69Table 4.16 Percentages of Achievement Tests administered in fivesubject-matter categories, by year, 1956-57 through 1968.69. 110'fable 4.17 Volumes for specific Achievement Tests expressed aspercentages of total Achievement. Test volumes, by year,1121956-57 through 1968.69'fable 4.18 Attendance at Achievement Test Losions, classified by113candidate's educational level, 1956-57 and 1968-69Table 4.19 Means, standard deviations, and numbers of cases (inthousands) for all candidates taking Achievement Tests,113by year, 1956-57 through 1968-69Chat,:. VTable 5.1 Correlation coefficients of selected faculty ratings with generalTable 5.2Table 5.'Table 5.4Table 5.5Table 5.6desirability rating and with first-year gradesExample of expectancy table prepared by Ile College BoardValidity Study ServiceExample of scatter diagram prepared by the College BoardValidity Study ServiceRange of correlation coefficients that would be expected toinclude 95 percent of ohserved values for selected populationvalues and sample sizesRelation between standing on predictor and standing oncriterion for various values of the correlative coefficientValidity coefficients of sAT and high school record: selectedpercentiles of validity coefficients based on students in liberalarts and general programs classified by sexTable 5.7 Validity coefficients of SAT and high school record: selei tetipercentiles of validity coefficients based on students in liberalarts and general programs classified by level of verbalability and by sexTable 5.8 Validity coefficients of SAT and high school record: selectedpercentiles of validity coefficients based on students in li'seralarts and general programs classified by homogeneity of verbalalJility and by ex122123125126127128129.r10120

Table 5.9 Validity coefficients of SAT and high school record: frequencydistributions and selected percentiles of validity coefficientsbase i on liberal arts and general groups classified by131homogeneity in verbal abilityTable 5.10 Validity coefficients of SAT and high school record: results for133engineering, science, and architecture groupsTable 5.11 Val dity coefficients of SAT and high school record: results for134bus:ness, education, and other groupsTable 5.12 EffIxt on multiple correlation cu-fficient of using AchievementTest scores with SAT and high school records for predictingfirst-term or first-year grades: women in liberal arts and136general programsTable 5.13 EPect on multiple correlation coefficient of using AchievementTest scores with SAT and high school record for predicting firsttei m or first-year grades: men in liberal arts and general137programs and engineering studentsTable 5.14 E feet on multiple correlation coefficient of using AchievementTest scores with sAT and high school record for pre3ic.ingfist-term or first-year average grades: various stud ant groups 138Table 5.15 Correlation coefficients of SAT-V, English Coinposit'On Test,140a Id high school record with English gradesTable 5.16 Correlation coefficients of sAT-M, Mathematics AchievementTests, and high school record with freshman matherlatics grades. 141Table 5.17 Correlation coefficients of SAT, Science and MathematicsAchievement Tests, and high schcol record with science grades. 142Table 5.18 Correlation coefficients 6, SAT-V, foreign language Achievement'I'ests, and high school record with foreign language grades. 143Table 5.19 Correlation coefficients of SAT, Social Studies Achievement143Test, and high school record with social studies gr idesChapter l'AThole 6.1Relative achievement of the CH A, CI1EM, and control groupsthree tests in chemistry177

ForewordThe College Entrance Examination Board has long since ceased to be engagedsolely with examinations, and the examinations it now sponsors are not all concernedwith college entrance. It is not surprising, however, that in the minds of many, theCollege Board is associated with "the College Boards"the Scholastic AptitudeTest and the Achievement Testswhich constitute the Board's AdmissionsTesting ProgramThis program, t largest by far of the Board's testing programs, has entered thelives of over a million students in each recent year, affecting the decisions made byhundreds of cclleges and the counseling services offered in thousands of schools. Itis natural that a program involving so many interests should have given rise tomany questions about its nature and value. The Board is indebted to Educational Testing Service (ETS), which administers the Admissions Testing Program,for providing in this book information that will answer questions about the moretechnical aspects of the program.It is a genuine pleasure for me, on behalf of the College Board, to thank mycolleagues on the staff of ETS who worked so hard and well to assemble and reportthese facts about the Scholastic Aptitude Test and Achievement Tests. I am keenlyaware of the special burdens that William Angoff has borne as editor, and sharewith all who were involved in the project admiration and appreciation for the graceand skill with which he brought the project to completion.The publication appears at a time when changes in the Board's testing programsare in the wind, some of them stimulated by the work and report of the Commissionon Tests, appointed in 1967 by Richard Pearson, former President of the Board,to review all of the Board's existing examinations. gather evidence of the need forchange, and consider what tests 'ray be needed a decade hence." This book showsthat significant changes have taken place time after time in the Board's collegeentrance tests, and it is fair to assume that further changes may well be necessaryif the program is to continue to fit the reds of col'iges, schools, and students. Theresourcefulness and talent underlying the techniques described in this book offerpromise that the best possible efforts will be brought to bear on the development ofwhatever new tests and techniques the needs of the future demand.Job/ A, Valoft(neCollege Entrance Examination BoardSeptember 197012

PrefaceThis repo,t, was prepared for the specialist who needs technical information aboutthe Admissions Testing Program of the College Entrance Examination Board. Itbrings together the findings of a considerable number of studies of the ScholasticAptitude Test (sAT) and the Achievement Tests. It assumes a reader fairly wellacquainted with the concepts and statistics of educational rreasurement.Three publications might be considered the book's predecessors: College BoardScores, Their Use and Interpretation, No, 1, published in 1953; College Board Scores,Their Use and Interpretation, No. 2, published in 1955; and 1957 Supplement toCollege Board Scores, No. 2. These publications are a compendium of availableinformation on the College Board Admissions Testipl Program up to 1957. Thepresent book occasionally refers to the earlier documents, but focuses mainly onmore recent data.The emphasis of this book is also somewhat different from that of its predecessors.The earlier publications were, in a sense, "how to do it" books. Their purpose wasto describe how the tests might properly be used in admissions and guidance andto provide data needed for interpreting the test scores. To a considerable extent,this function 1ms been taken over by College Board Score &ports: A Guide forCounselors and Ad missions Officers, which is revised annually and widely distributed.Mention should also be made of the computation manual, Predicting College Grades,published in 1961, and the Manual of Freshman Class Profiles, which was publisheduntil 1969, when it was incorporated in The College Handbook.These progeny of the old us,:r-andinterpretation books retain some of the "howto do it" emphasis of their ancestors, but tend to be mor interpretive than explicative. They do not, however, provide all the data needed for a comprehensivetechnical evaluation of the tests. The aim of this book is to fill that gap.The Board's Admissions Testing Program must respond effectively to a varietyof unique and complex demands, In order to indicate the setting within which thegrogram operates, the first chapter describes the program as a wholeits history,how it is organized, the unique demands it faces, and technic 11 inc,t ters involved inits operation that arise from the demands.The second chapter discusses the SAT, the third is about tl.e Achi vemerit Tests.The fcurth chapter interprets summary statist ics for various aubgroups of students,and the fifth reports on some of I he extensive validity studies of the admissionstests. The sixth and final charter summarizes the major findings of special studiesthat have been made in connection with the SAT and the Aciievernent Tests.Although there k quite properly. an interrelationship Z1/110114 the topics coveredin the various chapters. the drainers were written independently with the relentthat .arch would stand as an integral unit, containing all the ni,.terial relevant to itHowever, since some material, especially that having to do with technical pro13

cedures, has relevance to more than one chapter, there is inevitably some repetitionacross chapters.The authors were chosen for their special competence in the subject of theirchapters, but the book as a whole is the product of many other persons, most ofwhom are connected in one way or another with the Admissions Testing Program.In fact, the contributors include a substantial number of the staff members of theCollege Board and Educational Testing Service, and for this reason it would bepractically impossible to list them in this preface for the credit they richly merit.Especially 1.elpful were those individuals who served with me on the review com-mittee, some of whom were also chapter authors: William E. Coffman, John M.Duggan, Henry S. Dyer, William B. Schrader, Robert J. Solomon, and John A.Valentine.Additionally, the following members of the College Board Committee of Examiners in Aptitude Testing were of great aid in reviewing chapters in manuscript form:Carl t3ereiter, Frederick 13. Davis, John R. Hills, Julian C. Stanley (Chairman),Warren S. Torgerson, and Dean K. Whit la. A special debt, of gratitude is owed toDavid E. Loye for his able editorial ass:stance and to the staff of the ETS publicationsdivision for coordinating publication of the book.The Admissions Testing Program i3 revised continually to meet changing con-ditions,. Thus, it can be expected that this book, which describes the program,will be revised as necessary to report new developments in the program and toreflect the results of research on new as well as old problems.Witham If. AngoffEdLcational Testing ServiceSeptember 1970y111

(.:} Ii' FF,1 tThe Admissions Testing ProgramWILLIAM H. ANOOFF and HENRY S. DYERhistorical bockgroundThe College Entrance Examination Board was organized during a meeting held at Columbia University inNew York on November 17, 1900. The move culminated efforts over several years in this direction bythe Association of Colleges and Preparatory Schools ofthe INfiddle States and Maryland, and a number ofeducators in the East, notably Nicholas Murray Butlerof Columbia and Charles W. Eliot of Harvard. Theevent was, according to the Board's first historian, thefirst (organized "attempt to introduce law and olderinto an educational anarchy which toward the dos:, ofthe ninetecitli century had become exasverating, indeed intolerable, to sehoolamsters." tFuess, 1950, p. I)*At this time there was, in the opinion of the foundersthat a system of examinations of this sort would effectmarked savings of time, money, and effort in a(""ministering college admissions, that it would greatly aid thework of the secondary schools by reducing confusionand easing the strain on students, and that it wouldrepresent a cooperative effort of a group of colleges andsecondary schools to achieve a set of common goalswithout asking the colleges to surrender their prerogatives as to the particular examinations they requiredof their applicants or the manlier in which they mightwish to select their students.It is important to recognize that the examinationswere secondary to the main purpose of the Board,which was to provide a channel of communication be-of the Board, appallingly little agreement among thetween the schools and colleges and to encourage a degroe of uniformity in the secondary school curriculum.On the other hand, it is more tlian likely that these pur-colleges about the types of subject - matter pmpariotionand E tandards of proficiency required of the applicants.instrumentality of the examination system which, uponOne headmaster, for example, complained that out ofacceptance by the colleges as a way of setting standards,40 it r so boys preparing for college in his school, he had,effectively paved the way for the introduction of uni-in reality, more than 20 different classes. The effect ofform curriculums in the schools.this odiversity on the secondary schools was to make thetask of preparing their students for college an extremelydifficult one, cod made the task of the student whq wasIn its first year of op ertion the College Board heldessay examinations in nine subjects: English, French,German, Lit in, Greek, history, mathemat irs, chemistry,and physics. The definition of the requirements in eachsubject tatas taken from the recommendation of theprofessional association in each subject-matter area.not ,ure which college he hoped to attend difficult andconfusing as well.lio its atter Ipt then, to accomplish its 'impose andintroduce an order into the transition from school tocollege, the College Board established the beginningsof a system of syllabi or "course requirements" onposes would never have been achieved without theThe substance of tetath examination was then c'etcrminedby a carefully selected committee of examiners. consisting of well-known teachers and scholars in the lead-which schools and colleges could agree, and which mighting colleges and secondary sc.lools in the East, who metform the basis of a system of examinations offering theto confer on the gt neral content and structure of theunVormity that was so badly needed. That is to say,the examinations would be unicorn, in subject natterand uniformly administered at unolorm times, but heldexaminations and to decide on the specific questions tobe asked. After extensive preparation, the first exam-in many places to meet the convenience of tlic students,June 17, 1901, to 973 candidates at 69 testing centers.and they would be uniformly graded. It was expectedThus, although the Board was as yet far from Iing anational institution-75S of the 973 candidatel wereseeking admission to either Columbia or Barnardit*1' le authors wish to acknowledge e.-si,tanee i ro% hicd byti c outline, and, in many instanots, the particular pt, usingu,ed in Dr. Yuess*excellent oc FS kw ei the lion .crs history.They also wish to acknowledge the in% aluabie enures of int.( rmatk.n provided by the Culkge Board's Annual Reports.inations were finally cdmmistered during the week ofwas at luasi -a going operation with the beginnings of amechanism for continuation.As with the committees ( f examiners, committees ofreaders were also chosen with care. These committees,

one in each subject, were assembled in the ColumbiaUniversity Library and graded i he papers in accordancewith a proce

Table 2.3 Item types used in svr-verbal and SAT-mathematical sections, 1926-69. 21. Table 2.4 Classification scheme for sAT-verbal items. 22. Table 2.5 Distributions of SAT-verbal items by difficulty level. 21. Table 2.6 Distributions of sAT-mathematical item by difficulty level. 25. Table 2.7 Distribution of equated deltas for sAT-verbal and SAT-