If you have been researching the MAT, you have probably seen questions asking about the “Miller analogies test old scoring.” Let’s take a look at how the MAT used to be scored.
MAT old scoring
The Miller Analogies Test old scoring was based on the raw score instead of the scaled score. Prior to October 2004, the MAT consisted of 100 questions. Your raw score was the total number of questions you answered correctly. Similar to today, you were also given a percentile ranking that allowed you to compare your score to the other test takers during that time.
After Octorber 2004, the MAT was switched over to an entirely computer-based exam. The amount of test questions also changed. Currently, each MAT exam has 120 questions: 100 core that count towards your score, and 20 experimental questions that do not. The experimental questions most likely came about because Pearson needed to be able to create more versions of the MAT exam as the amount of test-takers increased.
The last big change is the switch from a raw score to a scaled score. At present, Pearson does not provide any sort of information on your raw score after you take the MAT. This is unfortunate since most students would like to see this information. Using a private algorithm, Pearson takes your raw score and converts it to a scaled score between 200 and 600.
If you’re wondering why they would do this whole conversion, it’s to better allow comparisons between MAT exams. Each form of the exam has a different set of questions; this leads to some forms being slightly harder or easier than other forms. The scaled score is supposed to be able to correct for these differences and allow fair comparisons between students.
A note on the old percentile ranks
Your percentile rank is based on a cohort group, a massive amount of test scores recorded during a set of several years. Pearson has updated the cohort group at least twice since 2004. This means anyone looking to compare their old percentile rank to the modern standards will not be able to do so with any sort of accuracy.