College Board Scholastic Aptitude Research Articles

Psychological profiles of the mathematically talented: some sex differences and evidence supporting their biological basis.

For over 20 years, above-level testing with the College Board Scholastic Aptitude Test (SAT) has been used to assess the abilities of well over 1,000,000 highly able 12-13-year-olds (students in the top 3% in intellectual ability). In this population, the predictive validity of the mathematical part of the SAT, SAT-M, for academic and vocational criteria has been demonstrated over 10-year gaps. Here, we document aspects of the psychological and achievement profiles of these highly able students, paying particular attention to sex differences. Males score higher on SAT-M (i.e., mathematical reasoning ability) than females; this difference is accompanied by differences between the sexes in spatial-mechanical reasoning abilities and in a number of lifestyle and vocational preferences. Collectively, these attributes appear to play a key role in structuring male-female disparities in pursuing advanced educational credentials and careers in the physical sciences. After profiling a number of the behavioural characteristics of the highly able, we examine some underlying biological correlates of these phenotypic manifestations. These include hormonal influences, medical and bodily conditions and enhanced right hemispheric activation.

METHODS USED TO ESTABLISH SCORE COMPARABILITY ON THE ENHANCED ACT ASSESSMENT AND THE SAT®

ABSTRACTRecently Marco and Abdel‐fattah (1991) reported newly established relationships between scores on the enhanced American College Testing Program (ACT) Assessment and scores on the College Board's Scholastic Aptitude Test (SAT). The current report provides a detailed description of the methodology that was used to develop the “concordance” tables reported in that study.Fourteen large universities provided data on applicants who had taken both the enhanced ACT Assessment and the SAT. The ACT Composite scores and the SAT‐verbal and SAT‐mathematical (SAT‐V + M) scores used in the comparability study came from ACT Assessment test editions administered from October 1989 to June 1990 and SAT editions administered from March 1989 to June 1990. The total sample consisted of 40,492 students. A subsample of 40,051 students who look the enhanced ACT Assessment and the SAT no more than 217 days apart was also used in the comparability study.The equipercentile procedure, a curvilinear procedure, was used to scale the scores. Steps were taken to ensure the accuracy of the conversions by (1) weighting the scores to reduce or eliminate the effect of the time differential between ACT and SAT testings or (2) smoothing the score distributions before scaling. The score conversions derived by applying the weighting procedures provided score adjustments that were in the direction implied by the length of time between testings.

ELIGIBILITY ISSUES AND COMPARABLE TIME LIMITS FOR DISABLED AND NONDISABLED SAT® EXAMINEES

This project used data from test administration timing records, the College Board's Scholastic Aptitude Test (SAT) history file, and a survey questionnaire to investigate two issues: comparable testing time and eligibility for special test accommodations for SAT examinees with disabilities. Comparable testing time for disabled examinees was found to be between 1½ and 2 times that for nondisabled examinees. Time limits in that range would assure that approximately equal percentages of disabled and nondisabled students would complete each section of the SAT The exception was blind students using braille or cassette versions of the test, who required between double and triple time. Eligibility for special test accommodations is tied to severity of disability and documentation of disability. Some levels of disability could be distinguished for students with sensory disabilities, based on self-reports and confirmed by the presence or absence of individualized education programs (IEPs). For example, IEPs were reported for 80 percent of totally blind candidates, 50 percent of legally blind candidates, and only 28 percent of visually impaired examinees who were not considered legally blind. Similarly, 90 percent of hearing-impaired students using American Sign Language (ASL) reported having IEPs, compared to 76 percent of those using total communication (English and sign) and 42 percent of those for whom English was the preferred communication mode. Severity of disability could not be defined for physically disabled or learning-disabled examinees, except on the basis of IEPs. However, it was difficult to isolate the need for test accommodations based on school practices (e.g., IEPs or special schools or classes), since examinees eligible for special test administrations were often tested under standard conditions. Alternatives to the current eligibility policy are discussed, including a change to school-based criteria and the use of individualized testing programs. A change to empirically derived testing times is also discussed.

SEX DIFFERENCES IN PROBLEM‐SOLVING STRATEGIES USED BY HIGH‐SCORING EXAMINEES ON THE SAT‐M

ABSTRACTAn item classification scheme developed by Gallagher (1990) was refined, resulting in a more accurate prediction of sex differences in performance on the mathematical sections of the College Board's Scholastic Aptitude Test (SAT). When Differential Item Functioning (DIF) procedures were performed on item data for examinees scoring above 650, the majority of items that were flagged as favoring males required the use of mathematical insight, whereas all the items flagged as favoring females required standard algorithmic solutions.Structured interviews were conducted with students (25 males and 22 females) in this score range to determine the nature of differences in strategy use. A classification scheme was developed for strategies that paralleled the item classification categories. There was substantial overlap in strategies used by males and females; however, analyses of strategy types used across all items indicated that females were more likely than males to use algorithmic strategies and males were more likely than females to use insightful strategies. It should be noted that these findings constitute a generalization across a group of subjects and items, and that there were several individual instances of males using more algorithmic strategies than females and females using more insightful strategies than males.Questionnaire data gathered from students who participated in interviews indicated a positive relationship for both males and females between SAT‐mathematical performance and positive attitudes toward mathematics (e.g., liking mathematics as a subject and recognizing its usefulness to their adult lives). The use of algorithmic strategies, however, was correlated with negative attitudes toward mathematics (e.g., mathematics being difficult and not being relevant to their lives). Implications of these and other findings are discussed.

SEX DIFFERENCES IN THE PERFORMANCE OF HIGH‐SCORING EXAMINEES ON THE SAT®‐M

ABSTRACTPerformance of high‐scoring males and females on the mathematics section of three forms of the College Board's Scholastic Aptitude Test (SAT‐M) was examined to determine how item content, solution strategy, and speededness differentially affect performance. The mathematical and verbal sections of the SAT were also compared for similarities in the performance patterns of high scorers.Items on three forms of the SAT‐M were classified on the basis of content and potential solution strategies. Male and female performance in the top 5 percent was examined using the Mantel‐Haenszel Differential Item Functioning (MH‐DIF) procedure. A main effect was found for item‐solution strategy, but not for item content. Males tended to outperform females on items requiring logical estimation or insight even when subjects were matched on years of mathematics course taking and average mathematics grades.Conventional measures indicated that the test was not differentially speeded. However, omit rates indicated that females omitted a greater proportion of items requiring estimation. Different patterns by gender were found on the mathematical and verbal sections of the test.

Long-term effects of acceleration on the social-emotional adjustment of mathematically precocious youths.

The Study of Mathematically Precocious Youth (SMPY) identified over 2,000 12-14 year-olds who scored as well as a random sample of high school females on the College Board Scholastic Aptitude Test. SMPY encouraged these students to accelerate their education; over 50% did. Their social development at age 18 and at age 23 was then assessed. We investigated the effects of amount and type of educational acceleration (grade skipping and subject matter) on psychosocial indices (self-esteem, locus of control, self-acceptance/identity, and social interaction)

EFFECTS OF COACHING ON THE VALIDITY OF THE SAT: A SIMULATION STUDY

ABSTRACT The effects of student coaching in preparation for the College Board Scholastic Aptitude Test (SAT) on the predictive validity of this test for freshman year performance were studied using data on 1985 freshman year students from four colleges. After the validity of the SAT was estimated for each school, a given proportion of students was picked, and a given magnitude of score gains was added to the observed SAT verbal and mathematics scores. The validity of the new SAT scores was then estimated and compared to the observed validity measures. Each simulation provides an estimate of the change in velidity resulting from a hypothetical coaching effect. Sample sizes from the 4 colleges were: (1) 996 students from a highly selective school; (2) 1,346 from a moderately selective school; (3) 386 from a small, less selective school; and (4) 203 from a small religiously affiliated college. Results indicated that the validity indicators of the SAT in highly selective colleges are more contaminated by coaching effects than in less selective colleges. Results also indicate that changes in the levels of coaching do not seem large enough to account for the observed decline in the validity of the SAT over the past 10 years. Appendix A provides a detailed discussion of the second college, and Appendix B describes decrements in multiple partial correlation coefficients. (Contains 5 figures, 14 tables, and 19 references.) (SLD)

Enhanced right hemisphere involvement during cognitive processing may relate to intellectual precocity

In the present study, intellectually precocious and average ability youths performed a dichotic listening task (Experiment 1) and a free-vision chimeric face task (Experiment 2). Patterns of hemispheric lateralization and the relative involvement of the left and right hemispheres during cognitive processing were assessed. In Experiment 1, the average ability youths demonstrated a right ear/left hemisphere (re/LH) superiority for identification of CV syllables, while the gifted subjects failed to show any ear/hemisphere advantage. In Experiment 2, members of both groups tended to judge the leftside smile/rightside neutral half-faces as “happier”, a pattern indicative of enhanced right hemisphere (RH) arousal. Notably, the degree of RH involvement was significantly greater in the gifted as compared to average ability youths. Moreover, laterality scores of the precocious on the chimeric face task correlated with their performance on the College Board Scholastic Aptitude test (SAT), i.e. the greater the leftward bias, the higher the SAT score. These findings, taken in composite, suggest that a high level of RH involvement during cognitive processing may be related to intellectual precocity.

EQUATING THE SCORES OF THE PRUEBA DE APTITUD ACADÉMICA™ AND THE SCHOLASTIC APTITUDE TEST®

ABSTRACTThe present study is a replication, in certain important respects, of an earlier study conducted by Angoff and Modu (1973) to develop algorithms for converting scores expressed on the College Board Scholastic Aptitude Test (sat) scale to scores expressed on the College Board Prueba de Aptitud Académica (paa) scale, and vice versa. Because the purpose and the design of the studies, though not all of the psychometric procedures, were identical in the two studies, the language of this report often duplicates that of the earlier study. The differences in procedure, however, are worth noting, and it is hoped that this study will contribute in substance and method to the solution of this important problem.The study described in this report was undertaken in an effort to establish score equivalences between two College Board tests—the Scholastic Aptitude Test (sat) and its Spanish‐language equivalent, the Prueba de Aptitud Académica (paa). The method involved two phases: (1) the selection of test items equally appropriate and useful for English‐ and Spanish‐speaking students for use as an anchor test in equating the two tests; and (2) the equating analysis itself. The first phase called for choosing a set of items in each of the two languages, translating each item into the other language, “back‐translating” independently into the original language, and comparing the twice‐translated versions with their originals. This process led to the adjustment of the translations in several instances and, in other instances, to the elimination of some items considered too difficult to be translated adequately. At this point both sets of “equivalent” items, each in its original language mode, were administered as pretests, chiefly to determine whether the two response functions for each item were sufficiently similar for the items to be considered equivalent.On the basis of these analyses two sets of items—one verbal and the other mathematical—were selected for use as anchor items for equating. These were administered again (in the appropriate language) at regularly scheduled administrations of the sat and the paa. An item response theory (irt) model was used to equate the paa to the sat, with the anchor items serving as the link in the equating process.The equating itself showed definite curvilinear relationships in both verbal and mathematical tests, indicating in this instance that both sections of the paa are easier than the corresponding sat sections. The results also showed good agreement between the current conversions and the 1973 Angoff‐Modu conversions for the mathematical tests, but not so close agreement for the verbal tests. The reasons for the difference are (speculatively) attributed to improved methodology in the present study, especially for the more difficult verbal equating, and to the possibility of scale drift in one or the other test (or both tests) over the intervening 12 to 15 years since the last study.

The Predictive Validity of a Study Habits Inventory for First Semester Undergraduates

In an effort to provide additional information concerning the relationship between study habits and college grade-point average (GPA), the College Adjustment and Study Skills Inventory (CASSI), high school rank (HSR), and the Verbal as well as the Mathematics scores on the College Board Scholastic Aptitude Test (SAT-V and SAT-M) were intercorrelated. Three stepwise multiple regression analyses were computed, with GPA as the criterion measure. One multiple regression equation used as predictors six subscales of the CASSI; another employed HSR, SAT-V and SAT-M; and the third, used a combination of HSR and the Time Distribution scale of the CASSI. Also, differences between means of males and females on all measures were calculated. Results showed that all the academic variables and all but one of the CASSI variables were correlated significantly with GPA; females were significantly higher than males on HSR and GPA; and males were significantly higher than females on SAT-M. The combination of SAT-V, SAT-M, HSR, and Time Distribution yielded a coefficient of multiple determination (R2) that was 16% higher than that obtained from HSR alone.

EELWORMS, BULLET HOLES, AND GERALDINE FERRARO

ABSTRACTThere is no safety in numbers. When data are gathered from a sample in which the selection criteria are unknown many problems can befall the unwary investigator. In this paper we explore some of these problems and discuss some solutions. Our principal example is drawn from data from students who choose to take the College Board's Scholastic Aptitude Test (the SAT). We explore methods of covariance adjustment as well as more explicitly model‐based adjustment methods. Among the latter we discuss Heckman's Selection Model, Rubin's Mixture Model, and Tukey's Simplified Selection Model.

SPECIFYING THE CHARACTERISTICS OF LINKING ITEMS USED FOR ITEM RESPONSE THEORY ITEM CALIBRATION1,2

ABSTRACTThe purpose of this study was to investigate the effects on IRT true‐score equating results of the characteristics of the linking items that are used to place parameter estimates for items contained in two test editions on the same scale prior to equating. The study was carried out using the three‐parameter logistic item response theory model and Monte Carlo procedures. So that the simulated data reflected actual test data, the true item parameters were taken from the estimated parameters obtained from LOGIST calibrations of item responses obtained from selected administrations of the verbal sections of the College Board Scholastic Aptitude Test (SAT‐V).The study was conducted in two phases, with the second phase based on results derived in the first phase. In the first phase, the effects of the characteristics and number of linking items were investigated using two of the common linking or scaling designs: concurrent calibration and the characteristic curve transformation method. The research extended previous research by systematically investigating the number of linking items needed to insure adequate equating results. In addition, the effects on equating of two different characteristics of the parameter estimates of the linking items were investigated: 1) items with parameter estimates having standard errors of estimation similar to those in typical SAT‐V common item sections, and 2) items with parameter estimates which have small standard errors of estimation. Finally, the effect on true‐score equating results of using peaked and uniform distributions of abilities for groups used to estimate the parameters for the items in the test editions to be equated was investigated.In the second phase of the study, the effects on equating results of common item sections that contained a few items whose item response functions were not well fit by the three‐parameter logistic model and of common item sections that contained a few items on which the two groups taking the common item section responded differently were systematically investigated. These effects were investigated using both the concurrent calibration and the characteristic curve transformation linking procedures for varying numbers of common items. The effects of these common item sections were studied using only, in most cases, a uniform ability distribution and characteristics of the parameter estimates selected as a result of the first phase of the study.The results of Phase I and II of the study indicate that, for data such as those simulated for this study, improved equating results are obtained if longer linking tests are used with a uniform distribution of examinee ability.The results of the study that investigated the quality of the linking item parameter estimates were confounded by the total test representation of the various linking tests. One possible conclusion from this aspect of the study might be that it is more important to evaluate the relative efficiency of the linking items, with respect to the total test, than it is to evaluate the size of the standard errors of estimation of the linking item parameter estimates.Finally, the study indicated that equating results, particularly those obtained using a characteristic curve transformation design, are affected by the presence of linking items that function differently for the two groups used to provide data for the equating and calibration. This is particularly true for shorter linking tests. It appears that the quality of an equating depends to some extent on prior screening of linking tests and removal of items that function differently for the two groups.

Selection of mathematically talented students

This paper reports a study concerned with the selection of mathematically talented students at the beginning of secondary school in Australia using a version of the College Board's Scholastic Aptitude Test (SAT). Some young students achieved scores comparable with older mathematically able students and with college-bound students in USA. Teacher nomination of students as talented resulted in the selection of considerably different groups of students to those selected using the SAT. Marked sex differences, in favour of boys, are noted and discussed. Three different procedures for comparing item responses of younger and older groups are employed, leading to the conclusion that younger and older students respond to SAT items in qualitatively different ways. Comparison procedures based on the Rasch model for measurement are favoured.

The Relationship of Laterality to Academic Aptitude

Self-report subscale scores from the Herrmann Brain Dominance Survey and Torrance and Reynolds' Your Style of Learning and Thinking were correlated with the College Board Scholastic Aptitude Test (SAT) Verbal (V), Mathematics (M), and Total (T) scores and with the grade point averages of 40 junior female occupational therapy students. Findings included (a) a strong relationship between the Herrmann right cerebral and total cerebral scores with the SAT V, M, and T scores, (b) modest relationships between the Herrmann limbic scores and the SAT M scores, (c) no meaningful relationships between the Torrance and Reynolds instrument scores and SAT scores, and (d) no consistent relationships between the measures of laterality and grade point average.

Student Perceptions as Predictors of Academic Performance in College Developmental Studies

The investigation examined (a) perceptions held by students of their preparation to do college-level work, (b) attitudes towards the Developmental Studies program and the university, and (c) knowledge of Developmental Studies as predictors of academic performance. In addition, the high school grade point average (GPA) and the College Board Scholastic Aptitude Test scores were employed as predictors in order to provide a standard by which to compare the predictive efficiency of the student-perception variables. The academic variables, in comparison with those associated with student perceptions, accounted for more variance in first-quarter GPA. Of the student-perception variables, those related to a student's estimation of preparation for college-level courses appeared to have the most direct relationships to first-quarter GPA.

SEX DIFFERENCES IN THE ACADEMIC PERFORMANCE OF SCHOLASTIC APTITUDE TEST TAKERS

ABSTRACTThe number of women college students has increased dramatically over the past fifteen years; in this same time period, the average College Board Scholastic Aptitude Test (SAT) scores for women have declined more than the scores for men. This study summarizes recent evidence concerning the academic performance of women and men by examining sex differences among (1) all SAT takers, (2) test takers grouped by anticipated major field of study, and (3) college freshman year courses and grades.Three major research questions guided this study: Are there consistent differences in the intellectual abilities of men and women? Do precollege admissions variables predict college performance with equal accuracy for women and men? Have the contents or structure of the SAT contributed to observed sex differences in performance on the test?Though the men and women who take the SAT differ in high school courses and grade point average, major field and career interests, and socioeconomic backgrounds, neither these differences nor research evidence on differential cognitive functioning were found sufficient to account for all of the observed sex differences in performance on the SAT. Some but not all of the male‐female differences in scores on the mathematics portion of the SAT were accounted for by differences in the depth and recency of course preparation. The observed‐first‐year college grades of women were slightly underpredicted by test scores and high school grades, but the magnitude of the underprediction was very small. While specific changes in the contents or structure of the SAT could not be associated with score changes, it seems likely that numerous small adjustments have played some part in recent shifts among the average test scores of women and men.Consistent with recent literature on sex differences in cognitive performance, this study concludes that sex‐related SAT differences are very small relative to the generally similar levels of performance by men and women, and that using both test scores and high school records to predict first‐year college grades continues to work reasonably well for both sexes.

Academic Excellence in an Early U. S. High School

This paper explores orientation toward academic excellence in high schools in United States during 19th century. Through a study of first 80 years of Central High School in Philadephia, I find several factors which encouraged academic excellence: a meritocratic ideology, voluntary enrollment, competitive selection of students and faculty, and structure of public education at time. In comparison, modern public high schools have only a limited commitment to merit, compulsory attendance, automatic acceptance of all students, less competitive selection of teachers, and a different structure of public education. In 1983 a series of high-level panels charged that education in United States had lost sight of its once dominant objective, pursuit of academic excellence. The gloomiest assessment came from national Commission on Excellence in Education (1983:5) which warned, the educational foundations of our society are presently being eroded by a rising tide of mediocrity that threatens our very future as a nation and a people. Blame for this decline in excellence has been placed at all levels of educational system: from elementary schools, which are accused of failing to ground students sufficiently in basic skills, to colleges, which are charged with failing to provide necessary degree of advanced training. But it is public high school which has received greatest critical attention, and for good reason. The cumulative academic deficits of lower levels of education become more visible in high school; and since latter is highest level of schooling at which attendance is nearly universal, these deficits affect a broader spectrum of population than do failings of colleges. Perhaps most frequently cited evidence of decline in achievement at U.S. high schools is 40 to 50 point drop in average scores on College Board's Scholastic Aptitude Test since early 1960s (Commission on Excellence, 1983:8). However, this is not most compelling proof that could be offered, since this decline in scores was accompanied by a nearly 200 percent increase in number of students enrolled in college during same period (U.S. Bureau of Census, 1982:Table 259). Thus, at least part of drop in test scores may be result of an expansion of educational opportunity rather than a reduction in student achievement. This same caveat applies to another favorite example of growing mediocrity in U.S. high schools. The Commission on Excellence (1983:8) noted that U.S. students scored low on achievement tests compared with students from other industrialized countries; but as Husen (1983) has pointed out, this difference is result of differences in selectivity of various systems of secondary schooling. In United States nearly everyone attends high school and threequarters of students graduate, while in European countries high school is limited to a small number of better students. If much of evidence about low levels of achievement among U.S. high school students is unconvincing, evidence about failure of U.S. high schools to promote academic achievement is stronger. As an institution, modern high school is less academic in orientation and less rigorously academic in application than it once was. The growth of electives in high school curriculum has permitted students to avoid more difficult academic courses, with result that only 31 percent of high school graduates took intermediate algebra, 13 percent took

Sex differences in mathematical reasoning ability: more facts.

Almost 40,000 selected seventh-grade students from the Middle Atlantic region of the United States took the College Board Scholastic Aptitude Test as part of the Johns Hopkins regional talent search in 1980, 1981, and 1982. A separate nationwide talent search was conducted in which any student under age 13 who was willing to take the test was eligible. The results obtained by both procedures establish that by age 13 a large sex difference in mathematical reasoning ability exists and that it is especially pronounced at the high end of the distribution: among students who scored greater than or equal to 700, boys outnumbered girls 13 to 1. Some hypothesized explanations of such differences were not supported by the data.

The Comparative Predictive Validities of High School Rank, the Ammons Quick Test, and Two Scholastic Aptitude Test Measures for a Sample of Black Female College Students

In the prediction of college grade-point averages (GPA) for minority college students, researchers often have reported relatively low predictive validity for both the Mathematics and Verbal sections of the College Board Scholastic Aptitude Test (SAT-M and SAT-V) in comparison with that obtained for high school class rank (HSR). This circumstance has suggested the need for alternative predictors to replace SAT-V and SAT-M as admissions indicators, particularly in the instance of minority college applicants. One such potential replacement was thought to be the Ammons Quick Test (AQT), which requires ony a short time to administer (10 to 12 minutes). By means of a multiple regression analysis with data from a sample of 63 specially admitted Black female undergraduates, the relationship between the optimally weighted composite of predictors (SAT-V, SAT-M, HSR and AQT) and the criterion (GPA) was evaluated. The HSR, SAT-V, SAT-M, and AQT afforded predictive validity coefficients of .52 (p < .001), .28 (p < .05), .22 (p > 05) and .19 (p > .05), respectively. In the stepwise procedure, there was an R2 of .36. HSR contributed .27 to the variance in the criterion measure, SAT-V added .08, AQT yielded only an additional .006, and SAT-M contributed virtually zero. The AQT did not appear to be a valid predictor. It can be anticipated that the large number of predictors relative to the smallness of the sample probably led to an inflated estimate of the contribution of each predictor variable to the variance in the criterion measure, although the relative importance of each predictor in all likelihood was approximated fairly accurately.

Development and Validation of a Foreign Language Attitude Scale

The present studies examined the validity of a new foreign language attitude scale (FLAS). One study compared the validity coefficients of the FLAS with those of ability predictors. The FLAS was correlated .25 with college foreign language (FL) grades, while being essentially orthogonal to most ability predictors. High school GPA was correlated .50 with FL grades. The correlations of FL grades with the total scores and subtest scores on the Modern Language Aptitude Test ranged from .12 to .44. College Board Scholastic Aptitude Verbal and Mathematics sub-test scores were not significantly related to FL grades. Stepwise regression analyses indicated that the FLAS made a significant contribution to the prediction of FL grades. In a second study with a different sample of college students, correlations between the FLAS and standardized measures of FL reading and listening proficiency were in the middle .40s. Although further research is needed, the present findings suggest that the moderate relationship between the FLAS and FL proficiency may be used to augment the validity in the prediction of success in FL courses.