MAT reliability refers to how consistent test scores are. MAT reliability can be measured by internal consistency and standard error of measurement. In this post we’ll learn how MAT reliability is proven!
MAT Reliability: Internal Consistency
MAT reliability is validated by internal consistency with the Kuder-Richardson formula. The Kuder Richardson formula considers the types and level of difficulty of the questions within a test based on the number of items in the test, the standard deviation of the total score, and the amount of candidates correctly answering each item. The result of this formula represents an average of all possible split-half reliability estimates.
The results can range from 0.00 to 1.00. 1.00 is a perfectly consistent score. For all MAT test forms used from 2008-2011 each test scored a 0.90 or higher. Judging from the Kuder-Richardson formula the MAT is highly reliable.
MAT Reliability: Standard Error of Measurement
Standard Error of Measurement estimates the amount that scores would probably vary if a candidate were tested repeatedly with the same test. Repeated testing always results in some variation. No single test event ever measures a candidate’s actual ability with complete accuracy. For test day tips check out this post on how to avoid getting bored during the MAT.
To account for this, the Standard Error of Measurement is used. Basically, the test-taker’s score is used to determine a score range in which his true score can be found. Thus, the lower the Standard Error of Measurement, the closer candidates’ test scores are to their actual ability, and the greater degree of certainty that the test scores are reliable.
The Miller Analogies Test uses these two statistical calculations, Internal Consistency and Standard Error of Measurement to ensure that the exam is both fair and reliable. Judging by these two statistical evaluations the MAT is highly reliable!