Introduction
Jennifer Kobrin, voicing the opinion of many educators, comments that students in the United States are earning high school diplomas but not leaving with the knowledge and skill needed to be successful in College (RN-30, 2007). Greene and Winters who measured college readiness in 2002, estimated that only 34% of high school students had the skills and qualifications necessary for college. The National Center for Educational Statistics measured a student by his grades in class, his rank, and his results in the National Educational Longitudinal study test scores and the SAT and ACT entrance marks (Berkner and Chavez, 1997). In the latest trend, many colleges use SAT to predict student’s likelihood of success at their institutions. In my essay, I am discussing how this test evolved over the last 106 years and the pros and cons of using this test, and how adequate it is to assess the fitness of the college-bound student. Having delved into various research literature on the subject, I am surprised that plenty of statistics have been derived on the possible skills and abilities of college-going students. How the test should be framed and what type of questions must be included to ensure a skillfully qualified citizen able to give of his best to society and humanity would be obvious from my discussion.
The SAT
The SAT Reasoning Test is a standardized test for College Admissions in the United States. The College Board, a non-profit organization, owns, publishes, and develops the SAT. Earlier it was developed by ETS (Educational Testing Service). ETS now administers the examination (The SAT Reasoning Test, Wikimedia). The SAT measures the critical thinking skills needed for academic success in college. It is taken by high school juniors and seniors. The Board states that assessing these students based on the SAT in combination with high school grades provides a better assessment of possible success. It assesses how well test takers analyze and solve problems using the skills they have acquired in high school (The SAT Reasoning Test, Wikimedia) Studies over the years have spoken favorably for the SAT.
Many colleges use SAT scores and High School Grade Point averages to predict the student’s chances of success at college. According to the National Center for Educational Statistics (Berkner and Chavez, 1997), students were rated on a 5 point scale. They showed that 65% of students were eligible for college. Greene and Winters (2005) opined that the method had flaws as the rating was done on the student’s Grade Point Average alone. Other criteria for college readiness were neglected. Neither group considered actual student performance (Kobrin,2007). The American College Testing Programme (2004) indicated that they used
actual student performance in their ACT Assessment when reporting on college readiness benchmarks. The ACT showed that only 22% of students could meet the benchmarks in algebra, composition, and biology. Only 22% were ready for college.
SAT Reasoning Test to predict college readiness.
SAT scores and college grades of 165781 students who had valid SAT scores and GPAs who entered college in 1995 were used for analysis by Kobrin in 2007. Logistic regression was used to determine the predicted probability of success for each student, where success was defined as achieving a first-year cumulative GPA of 2.7 or higher, or 2.0 or higher. “The benchmark scores were computed for the total sample and separate benchmarks were computed within each of the 41 institutions. The mean institution-level benchmark weighted by the number of students at each institution, as well as the median benchmark across institutions, as compared with the benchmark scores based on the total sample.” (Kobrin, 2007). It was concluded that only 22% of students were ready for college in 1995 just as the ACT reported and in 2005, this became 25%. Males achieved a higher benchmark. The difference between genders was smaller in the verbal score benchmark than the mathematics score benchmark or the total score benchmark. The percentage of African American students were reaching the benchmark was lower. The Asian American students were higher. There are several ways to determine college readiness benchmarks. Results would vary with each. Colleges are thereby encouraged to find their benchmark scores based on their individual needs and to periodically revalidate with current data (Morgan and Michaelides, 2005).
Evolution of the SAT
Carl Brigham developed this test when he was one of the psychologists who worked on the Army Alpha and Beta Tests (The SAT Reasoning Test, Wikimedia). The test was originally developed in such a manner as to eliminate bias between people from different socio-economic situations. The College Board was set up in 1901 and the first test was conducted in 67 places in the US and 2 in Europe. It was not a multiple-choice test but evaluated based on essay responses.
In 1926, the first SAT administration was set up. The test was known as the Scholastic Aptitude Test. It had 315 questions on definitions, arithmetic, classification, artificial language, antonyms, number series, analogies, logical inference, and paragraph reading to be completed within 90 minutes (The SAT Reasoning Test, Wikimedia). In 1928, the time limit was increased and changes in the sections were made. Maths was entirely discarded leaving only the tests for verbal ability. In1930, the SAT was split into 2 sections of verbal and maths. This style continued till 2004. In 1936, analogies were reinstituted and the test for verbal ability lasted up to 115 minutes to answer 250 questions. The maths test had 100 free response questions to be answered in 80 minutes. Again the Maths section was eliminated from 1936 to 1941 (The SAT Reasoning Test, Wikimedia). When this maths section was readded in 1942, it had multiple-choice questions. 1946 saw the elimination of the paragraph reading from the verbal section. Reading comprehensions and sentence completions were added. Till 1957, the test time was 90 to 100 minutes for the approximately 150 questions in the verbal section. Till 1975, the time limit was steady at 75 minutes for 90 questions (The SAT Reasoning Test, Wikimedia). In 1954 questions on data sufficiency were added to the maths section and in 1974, this was replaced by quantitative comparisons. In 1974 the time limit for each of the verbal and maths sections was 60 minutes and thereby questions were reduced too.
The ETS studied the students from minority groups and low socioeconomic backgrounds, the Strivers Score Study (The SAT Reasoning Test, Wikimedia). In the research phase of the study from 1980 to 1994, a student identified from a minority or low socioeconomic background was awarded an additional 10-200 points depending on his race and gender. This was aimed at helping students from these groups to reach even the Ivy League College. When the Strivers Project became known to the public in 1992, it had to be terminated in the public interest. The Federal Courts looked into the matter and advised the authorities to alter their data collection process to include Strivers by considering only age, race, and zip code. These new changes have been effective since 1994 (The SAT Reasoning Test, Wikimedia ).
1994 saw some dramatic changes in the questions. Antonym questions were removed and more paragraph reading was instituted. Due to pressure from the National College of Teachers of Mathematics, a few non-multiple-choice questions were added permitting the students to give their answers. Calculators were introduced (The SAT Reasoning Test, Wikimedia). Simultaneously concepts of probability, slope, elementary statistics, counting problems, median, and mode were introduced. The average score was 1000.
In 2005 more changes were made due to criticism by the University of California. The New Sat came into being. The questions on analogies in the verbal section and quantitative comparisons in mathematics were removed. An essay was added to test the writing ability of the student. The New SAT or SAT Reasoning Test was first offered on March 12, 2005. The mathematics section began to cover 3 years of high school maths. The verbal section began to be called the Critical reading section (The SAT Reasoning Test, Wikimedia).
Structure of the SAT Reasoning Test 2005
There are 3 major sections: Mathematics, Critical Reading, and Writing. The addition of the writing section is a noteworthy improvement (Kobrin et al, 2006). Each section has scored in multiples of 10 from 200-800 (The SAT Reasoning Test, Wikimedia). Total scores are obtained by adding marks for all three. There are 10 subsections and an experimental section where marks are not awarded towards the final score. This experimental section which could be in any of the 3 has been included to normalize questions for future use by the administration. The actual time for the test is 3 hours and forty-five minutes.
General reliability and validity findings
Kobrin et al (January 2007) researched the comparison of the old SAT and the new one. When the SAT was revised in March 2005 to strengthen its alliance with the curriculum and instructional practices in college and high school, it was assumed that scores on the new Test would be fully comparable to and interchangeable with previous scores on the earlier SAT (Kobrin,2007). This was essential for reassuring test users to track score trends and also to enable colleges to treat all results equally when selecting their students. The two versions of the same test were measuring the same constructs (Angoff, 1971, 1984). Construct
comparability and equitability are intertwined topics. Dorans (2000) described 3 levels of score linkage, equating, scaling, and prediction ranging from strict exchangeability (equating) to a mere association between the scores. Exchangeability means that the scores from the two tests are interchangeable. A test taker who takes the two tests would obtain the same score on both after equating both. Another point to note is that a test taker who takes the same test twice can have different scores. However, the results of SAT and ACT are not exchangeable. Lord (1980) pointed out 4 requirements for two tests to be equated. These include construct comparability
(2 tests must have the same construct), subpopulation invariance (the equating transformation should be invariant across the subpopulations) and equal reliability. His equal reliability requirement is consistent with the ideas of Dorans and Holland (2000) and Angoff (1971,1984). A heavy responsibility rests on the shoulders of the test developer to equate the two tests.
By revision of the SAT in 2005, the overall constructs measured by the test had not changed. Field trial results showed that that the new critical reading and mathematics sections had similar reliability and standard errors of measurement to the prior verbal and mathematics sections. The correlation between the scores on the new critical reading and old verbal was 0.91 and the correlation between the scores on the new and old maths sections was 0.92. The speed involved in finishing both the old and the new tests was the same. The elimination of analogies had not influenced the construct of the test. Similarly, the addition of 3rd-year maths material did not influence the construct. Dorans and Holland (2000) consider that subpopulation invariance is the most important requirement in equating two tests. Angoff (1971, 1984) opined that the construct of the test and subpopulation invariance go hand in hand. This can be interpreted by saying that for 2 tests that have different constructs, the equating functions would be different for different groups. Research has not confirmed that racial or ethnic groups have invariance but it has been associated with gender.
Studies were conducted by John W. Young with the help of Jennifer Kobrin to evaluate the results of 49 earlier studies of differences invalidity/prediction and college admission testing They used first-year grade point average (FGPA) as the criterion and test scores (usually SAT scores) and high school grades as predictor variables in a multiple regression analysis. This study focused on racial differences and sex differences. Correlation coefficients were also usually reported as evidence of predictive validity. The first main conclusion that can be drawn from this review of research is that group differences do occur invalidity and prediction. Secondly, these differences varied considerably depending upon the group of interest. Thirdly group differences have not remained fixed and can change with time. Finally, the conclusion said that the major causes of group differences invalidity/prediction studies are not yet well known or understood.
SAT to identify talents
Since 1972, SAT has been used to identify intellectually talented 7th and 8th graders in high school (Lubinski et al, 2003-2004). This was done to facilitate the pathway of these talented young ones to a highly successful future. Graduate students (GS) and talent search (TS) students (who were identified in1980) were compared to each other by assessing their educational achievements several years later in 2003-2004. 51.7% males and 54.3% females of the talent search students earned doctoral-level degrees while 79.7% of the male graduate students and 77.1% of the females earned Doctor degrees. It was surprising to find that the differences between the 2 groups were not marked. The institutions at which the Talent search students reached were highly ranked. The female TS students secured coveted positions in good Universities. That the SAT can identify young children who attain prestigious positions at top universities is remarkable (Lubinski, David, et al). The SAT assesses much more than book learning. “Instruments such as the SAT assess much more than book-learning potential; they capture important individual differences in human capital critical for advancing and maintaining society in the information age through a variety of demanding professions, including medicine, finance, and the professoriate.”(Lubinski et al, 2003-2004).
The Calculator and SAT
Since 1994, calculators have been brought to the Examination hall for SAT. The majority of students used them for less than half of the questions. Girls used them more often. 96% of whites and Asian Americans brought calculators compared to 88% of African Americans and 90% of Hispanics (Scheuneman and Camara, 2002). Students who used calculators for 1/3 to ½
of the questions only fared better. Students who used calculators for more than half of the questions did not do well. Students with graphing calculators performed better than those with scientific calculators and those with four-function calculators performed worst.
Extended Time for SAT
The number of students with disabilities requiring accommodation on the SAT has increased. Most of these disabled children had learning disabilities (90%). The accommodations were varied. Presentation format may be in the form of Braille, cassette versions, large print forms, or a reader for the visually impaired. Response format may vary from oral responses to an aide or large bubbled answer sheets (Camara, 1998). An individual administration may be allowed for a test set. Timing may be changed to allow more breaks. Some students want more accommodations. However, usually, they only want more time. The extra time is permitted to allow the disabled students to compensate for the disability. For administrative efficiency of the SAT, 80% of examinees are to reach the last question or all examinees should finish 75% of questions. So the extra time allowed is justified. In the research, 4 groups of students were studied. Normal students and those with learning disabilities but with grants of extra time were able to score well. Only those students who availed more accommodations than the others scored much less even with the privileges. Students with lower scores on their first SAT usually scored higher with extra time in the next.
The SAT is considered as a power test but the time factor is downplayed as test speededness gives rise to unanswered questions, preventing real assessment of the student’s abilities. Speed of performance is not a key component. A study by Bridgeman and Cline was aimed at enabling students to finish a 25-minute section within 40 minutes. Certain sections of the test were given extra time for completion. The timing conditions were evaluated with an analysis of variances. Gender, ethnicity, and ability were taken into consideration. Extra time provided no advantage in the New Reading Section. Scores were slightly higher by less than half a formula score point in the New Mathematics section. Extra time produced a noticeable advantage in the writing scores by 1.4 formula score points higher. Differences were negligible in the higher and lower ability groups. The middle ability group showed an advantage which translated to about 28 points on a test. It is inferred that the high-ability group could have no advantage with the extra time. There is a chance that the extra time would cause the students to look through answers and possibly several correct answers would be changed to incorrect ones.
The conclusion was that the extra time was advisable.
Sections of the SAT
The critical reading section was formerly the verbal section. There are 2 twenty five minute sections and 1 twenty minute sections in this portion. Varying questions including sentence completions and questions on short and long reading passages are found. (The SAT Reasoning Test, Wikimedia). The student’s vocabulary, sentence construction, and organization are tested here.
The Mathematics section also has 3 portions: 2 twenty-five-minute and one twenty-minute portion. Most of the questions are multiple-choice questions. Straightforward symbolic or numeric answers only are expected (The SAT Reasoning Test, Wikimedia).
The writing section tests the student’s ability and knowledge of the English language.
It is a new section and has increased the credibility of the SAT Reasoning Test.
The Writing Section
The writing section has multiple choice questions and an essay. The multiple-choice questions assess how well students use standard written English and test the student’s ability to identify sentence errors, improve sentences and improve paragraphs (Kobrin et al, 2006). They also assess whether the student can use language that is consistent in tenses and pronouns. The student’s understanding of parallelism, noun agreement, and subject-verb agreement are assessed. Whether he can express ideas logically is also assessed. He needs to avoid vague and ambiguous pronouns, wordiness, and improper language (Kobrin et al, 2006). The error identification and sentence improvement questions test the student’s grammar. The paragraph improvement questions test the student’s understanding of the logical organization of ideas. The essay which is the first on the test will be having a philosophical base and will cover topics that do not affect their performance from whatever background they come from. 25 minutes are allotted for the essay. The essay score is roughly 30% of the marks and the multiple-choice questions carry 70% (The SAT Reasoning Test, Wikimedia). SAT essay prompts are developed according to some guidelines to allow a maximum number of students to write good essays.
Breland et al (2004) studied the impression of the essays on ethnic, language, and gender groups.
They could not find any negative impact. No group was disadvantaged. The scores are weighted equally. Each correct answer gets one point. An incorrect answer loses a quarter-point. For incorrect Mathematics grid-in questions, no marks are deducted. The examinations are conducted 7 times a year in the US and 6 times in other countries.
The goal of revising the SAT was to strengthen the links between the skills measured
by the SAT Reasoning Test, high school and College Curricula, and instructional practice (Kobrin et al, 2006). Several groups of high school and college educators helped to develop
and define the writing section.
Is there room for improvement in SAT?
Most of the skills considered important by teachers have been included in the SAT. However not all the skills deemed important have been included. (Kobrin et al, 2006). The omissions are creative writing, using peer groups for feedback and revision, responding to the needs of different audiences, using prewriting techniques to generate text, generating multiple drafts while creating and completing texts, understand writing as a process of invention and rethinking, learning strategies for revising, editing and proofreading, understanding the purposes of different kinds of writing and writing analyses and evaluation of texts. The inclusion of the essay is an important component for assessing writing skills. It also ensures the seven skills most required in the classroom. Only 2 common skills are not handled in the SAT: controlling spelling errors and using punctuation appropriately (Kobrin et al.2006).
Who values the essays?
The essays are scored by two independent qualified readers who would mark on a scale of 1-6 (Kobrin et al, 2006). If the scores of the 2 differ by more than one point, a third reader does the marking. The essays are transferred to the readers via the Internet. By working with the readers via the Web, a large reader pool is made. The qualified reader must have a bachelor’s degree at the minimum, teach at a high school or college-level course that requires writing, or have taught during the last 5 years. They should have taught for a minimum period of 3 years. Readers should be residing in the United States, Alaska, or Hawaii. They should be residents or citizens and be authorized to work in the US. The readers are also required to undergo a rigorous training course on the Internet. The training does not stop once selected. It is an ongoing process. The caliber of the readers is checked frequently (Kobrin et al, 2006). To check reader accuracy, validity papers are mixed with the student responses. Web-based scoring enables leaders to monitor the readers. This monitoring and training program helps maintain the high quality of scoring (Kobrin et al, 2006).
Information for the candidates
The candidates may sit for the whole test or any of the subjects. US Students have to pay $43 while international students pay $68. Students receive their scores 3 weeks after the test.
Each section carries 200-800 marks and the subscores for the essay and multiple-choice portion.
Students will receive their percentile too in comparison to the other test-takers. For a fee, students will receive the correct answers to the questions and an online explanation for each answer.
Technical information on the SAT Writing Test
This section of my essay presents the statistical and psychometric information on the first 6 administrations of the SAT Writing test (Kobrin et al, 2006). Annually, technical information on the SAT Test based on the scores of the previous years’ set of college-bound seniors is published. The report of 2005 would show the most recent SAT scores of the college-bound seniors of 2005. However, the aggregate report of the writing section and the percentiles would be included in the next years’ reports. As many students do not complete the SAT at one go, they would be finishing the sections including the writing section by the next year. When all the students have finished the writing section, data will be analyzed and reported in form of aggregate data and percentiles. However, there is an interim percentile table to guide in interpreting the writing score. These figures would change when the full information is available.
Findings in 2005
1258016 students took the SAT Test. The writing section’s average score was 502 with a standard deviation of 108. The average multiple-choice score was 50.2 with a standard deviation of 10.9. The distribution of essay scores was normal with a mean score of 7.3, a median score of 7.0 and a modal score of 8.0 (Kobrin et al, 2006). The essay reader agreement statistics show that they agreed with 56% of the essays. There was a one-point difference in scores for 38% of the essays and a two-point difference for 6% of them. Less than a half % showed a difference of three points.
Reliability of the writing section
The scores for the writing section correlated 0.84 with the critical reading score and 0.72 with the mathematics score. The inference is that there is a high degree of correspondence among the three sections of the SAT (Kobrin et al, 2006)
The internal consistency reliability estimates for the first seven forms in 2005 for the multiple-choice questions ranged from 0.88 to 0.9 with a mean of 0.89 across forms, based on the 49 items. These reliability estimates are similar to the estimates of the earlier SAT Subject Test in writing (0.86 to 0.92) (Kobrin et al, 2006).
A special study was conducted to estimate the reliability of the essay component (Allspach and Walker, 2005). The same student wrote two essays at an interval of two weeks. 3500 students from 45 schools participated. The training was similarly imparted to the raters as that given to the raters of the operational SAT essay responses. The inter-rater reliability, the essay-scoring reliability, and the essay-observed scoring reliability were determined. “The inter-rater reliability is the correlation between scores from
two raters scoring the same essay. The essay scoring reliability is the correlation between average scores from two sets of raters scoring the same essay. This represents the consistency in the scoring method itself. The Essay-observed score reliability is the correlation between average scores from two sets of raters scoring different essays. This represents the proportion of true score variance in the essay score itself and is the relevant coefficient to the use of essay scores to estimate examinees’ writing ability” (Kobrin et al, 2006). The estimated true score reliability across all prompts was 0.76. 4 prompts were used to study the standard error of measurement for the writing section. The average alternate-forms reliability coefficient across these prompts was found to be 0.6679. Based on these, the standard error of measurement for writing was found to be 1.04. This is interpreted as follows: “for a subscore of 8, there is a 68 percent probability that the student’s true score is between 6.96 and 9.04” (Kobrin et al, 2006).
Predictive Validity of the Writing Section
The American Institutes for Research was commissioned by the College Board to investigate the validity of the SAT Writing Section for predicting FGPA (first-year college grade point average) and course grades in college English Composition Courses (Norris, Oppler, Kuang, Day and Adams ). The focus was on the extent to which the new writing section predicted college performance taking into consideration the SAT Verbal and Mathematics scores and High School Grade Point Average (HSGPA). For this study, incoming first-year students in the fall semester of 2003 from 13 colleges and universities in the United States were given an experimental version of the new SAT (Kobrin et al, 2006). The participants numbered 1572 who had already been identified by the SAT Reasoning Test. Correlations were computed between each predictor (constituted by the SAT Verbal, Maths, Writing scores, and High School Grade Point Average) and each criterion variable (First Year College Grade Point Average and the English Composition Grade Point Average).
Hierarchical regression analyses were also done to assess the incremental validity results for the various combinations of predictors and FGPA. Overall study estimates were also obtained. Statistical procedures to correct for multivariate range restriction (Lord and Novick, 1968) and shrinkage (Rozeboom, 1978) were applied. The average corrected validity coefficient for FGPA ranged from 0.2 for the essay component to 0.51 for the SAT combined score. When corrected for range restriction, all the predictors except the essay had validity coefficients of 0.43 and more. The average corrected validity coefficient ranged from 0.18 for the essay to 0.35 for the HSGPA (Kobrin et al, 2006). The combined writing section, multiple-choice, and verbal scores were also fairly predictive of the ECGPA with corrected validity coefficients of 0.32,0.31, and 0.3 respectively. Correcting for range restriction and shrinkage, the incremental validity of the writing scores added to the verbal scores, the mathematics scores, and the HSGPA come to a total of 0.01. The multiple correlation for this fully corrected model was 0.6 and is considered the best estimate for predicting FGPA (Kobrin et al, 2006)
Conclusion
Assessment of college readiness is an essential component of Education. The evolution of a test over 107 years has seen changes as per the needs of the then society and the whims of the Educationists. It started with the name of Scholastic Aptitude Test in1901 and has reached the SAT Reasoning Test in 2005 with the SAT not particularly standing for any word.
Starting from essays questions, it has reached multiple-choice questions. The essay is added for testing the prowess in English essay writing only. Mathematics was ‘added and subtracted’ at intervals. Then quantitative comparisons were added at some time and eliminated at some other time. Antonyms went in and out similarly. The time limits and the number of questions kept
changing. The students were permitted to use the calculator at some time. The format of the SAT kept changing and now in 2005 has seen the latest changes. The writing section is added and has made the test better for the time being. Despite it not being a flawless assessment of the high school student’s readiness for college, it appears to be the best now. More changes are expected.
References
Allspach, J. R., & Walker, M. E. (2005). Essay Reliability Coefficients. ETS Memorandum, 2005.
Angoff, W.H. (1984). Scales, norms, and equivalent scores. Princeton, NJ: Educational Testing Service. (Reprinted from Educational Measurement [2nd ed.], by R.L. Thorndike, Ed., 1971, Washington, DC: American Council on Education.)
Berkner, L., & Chavez, L. (1997). Access to Postsecondary Education for the 1992 High School Graduates. (NCES 98-105). Washington, DC: U.S. Department of Education, National Center for Education Statistics.
Breland, H., Kubota, M., Nickerson, K., Trapani, C., & Walker, M. (2004). New SAT writing prompt study: Analyses of group impact and reliability. (College Board Research Report No. 2004-1). New York: The College Board.
Dorans, N.J. (2000). Distinctions among classes of linkages (College Board Research Note RN-11). New York: The College Board.
Dorans, N.J., & Holland, P.W. (2000). Population invariance and the equatability of tests: Basic theory and the linear case. Journal of Educational Measurement, 37(4), 281–306.
Kobrin, Jennifer K. et al, “Comparability of Scores on the New and Prior Versions of the SAT Reasoning Test™”, ( College Board Research Note,RN-31) 2007, New York, The College Board.
Kobrin, Jennifer L., “Determining SAT Benchmarks for College Readiness” ( College Board Research Note, RN-3o) 2007, New York, The College Board.
Lord, F.M. (1980). Applications of item response theory to practical testing problems. Mahwah, NJ: Lawrence Erlbaum Associates.
Lord, F. M., & Novick, M. R. (1968). Statistical theories of mental test scores. Reading, MA: Addison-Wesley.
Lubinski, David et al, Department of Psychology and Human Development, Vanderbilt University“Tracking Exceptional Human capital over two decades” 2003-2004.
Morgan, D.L., & Michaelides, M.P. (2005). Setting cut scores for college placement. (College Board Research Report No.2005-9). New York: The College Board.
Norris, Oppler, Kuang, Day and Adams, 2005 “Assessing Writing” Volume 10, Issue 3, 2005, Pages 151-156.
Rozeboom, W. W. (1978). Estimation of cross-validated multiple correlation: A clarification. Psychological Bulletin, 85(6), 1348–51.
Scheuneman, Janice; Camara, Wayne J.; “Calculator Use and the SAT I Maths”, (College Board Research Note, RN-16), 2002, New York, The College Board.
The SAT Reasoning Test,Web.
Wayne, Camara; 2000, “Testing with extended time on the SAT I :Effects for students with learning disabilities” (College Board Research Note, RN-08 ) 2000, New York, The College Board.