The Miller Analogies Test (MAT), published by Harcourt Assessment, Inc., measures analytic ability. The candidate must solve problems presented in the form of analogies. Analogies are relationships. The MAT tests the ability to recognize relationships between concepts, English language facility, and general knowledge. A basic knowledge of the natural sciences, social sciences, mathematics, and subjects in the liberal arts is required to write the test successfully.
The Miller Analogies Test is composed of 120 questions. One-hundred of these questions count toward the test score, and 20 are experimental questions. These experimental questions are being tested for use on future versions of the test. The candidate cannot opt out of answering these questions, however. The experimental questions are interspersed with the test questions, and the candidate has no way of knowing which are questions will count towards his or her score. The candidate is not advised to guess at which questions will count.
The candidate can opt for a computer-based test or a paper-and-pencil test. The MAT is published in a number of equivalent versions in both of these formats. The MAT does not come in foreign-language versions. The use of test aids, such as dictionaries, notes, calculators, or electronic devices is not permitted.
Many graduate schools use MAT scores in graduate school admission decisions because the MAT tests high-level analytical reasoning ability. Research has found MAT scores to be an excellent measure of analytical thinking and verbal understanding. Studies show a correlation between high MAT scores and success in academic and professional endeavors. MAT scores can help universities in admission decisions by identifying applicants of superior cognitive ability. The skills tested by the MAT are invaluable for success in graduate school. The MAT has been used for over fifty years and has been tested extensively for reliability and validity. Research conducted by Harcourt Assessment, Inc. and research conducted by independent researchers has demonstrated that MAT scores are predictive of success in graduate school.
Harcourt Assessment, Inc. continues to improve the MAT by investigating trends in responses and is continually creating new versions of the test. Twenty questions on each MAT administered to actual candidates are experimental questions used not in generating a score for the candidate, but to aid in the development of new versions of the test. By writing the MAT, candidates are giving permission for their results to be used in research.