Methodological Assessment of The SAT's
Format and Norming
Performance on The College Board Scholastic Aptitude Test is used as one criterion upon which college admission boards accept incoming students. The SAT is administered in six 30 minute sections (two verbal, two mathematical, one section for the TSWE which measures correct use of English, and one section for gathering research and constructing future test items) and measures three aptitudes. Items included antonyms, analogies, sentence completion, reading comprehension, mathematics items, quantitative comparisons, and language usage and sentence correction. It is multiple choice and measures verbal aptitude and mathematic aptitude on a 200-800 point scale and English usage on a 0 to 60 point scale.
It was normed with all demographic sub-divisions amongst high school students contemplating pursuing post-secondary education and new tests are consistently analyzed for antiquated or biased items.
Reliability and Consistency
Internal consistency is one of the primary concerns of the test. The K-R20 is typically in excessive of .90. Test-retest correlations are about.87 for both the math and verbal sections. Within the verbal sections the sub-scores vocabulary and reading have reliability coefficient of .85. The TSWE has a test-retest coefficient of .82.
The validity for the SATs is Criterion based (predictive) in that the Sat’s are supposed to help anticipate college success levels. Since 1964 more than 3,500 studies at 750 colleges have shown that SAT performance indeed correlates with college grades. This shows high criterion validity. Studies concerning content and construct validity have likewise been performed and it has been found, for instance, that the verbal section is reflective of language used in a typical college academic setting.
Although this review gave no specific citations of studies of validity this is probably because as the reviewer write, “The SAT remains the documented instrument of its kind,” and because such a, “multitude of studies concerning its reliability and validity have been and continue to be conducted.” Both the validity and reliability of this test are well-established and a specific method of test construction is used in developing consistency between each edition and all other editions. Strict administration procedures are uniformly followed independent of the administration site. For these reasons and for its strong predictive criterion reliability in such an important area of inquiry I would rate the test very highly.