Can MAT Predict Graduate School Performance?

“Can the Miller Analogies Test Predict Graduate School Performance?” It’s a good question. The answer, unfortunately, comes with a lot of qualifiers. If you’re very interested, you can read through the MAT reliability research. Otherwise, here is a quick overview.

What can the MAT predict?

The point of exams like the MAT is to predict how well you will do in graduate school. One of the most common things to look for is a correlation between MAT scores and first year graduate GPA. What we want to see is a positive correlation. You want to see students who score well on the MAT later achieve good GPAs in graduate school. That would hint at a positive correlation.

According to the research featured in the above link, there is a significant positive correlation between MAT scores and graduate school GPA. Unless you are a statistics student, that probably does not mean much to you. Let’s make some sense of it.

“Significant” in statistics just means “real.” Seriously.

Can Miller Analogies Test Predict Graduate School Performance?

When doing math calculations of this kind, it’s possible to end up with differences that do not represent the real world. Saying something is “significant” in statistics essentially means that we know what we are looking at is real. Not necessarily important. Not necessarily consequential. Just real.

A positive correlation means that a set of things appear together and they move in the same direction. If there is more of one, there is more the other; if there is less of one, there is less of the other. Back to the MAT.

The research shows that the MAT has a significant (read: real) positive correlation with GPA. Higher scores on the MAT tend to predict higher GPAs. However, correlations can be ranked on the basis of strength. On a scale from 0-1, the MAT averages between 0.37-0.41. This is considered a weak to moderate correlation.

But what does that mean?!

In short, I made you read through all this mathy-talk to say that the MAT is an okay predictor. Roughly speaking, a little less than half the students who score well on the MAT are going to achieve good GPAs. The rest of the time is a mixed bag: many students who received low MAT scores will still go on to attain higher GPAs. Because the correlation is only moderate, it simply does not predict all cases.

These exams are more about lowering risk for graduate schools. Instead of having no idea how a student will do, they know that students who score well on the MAT have almost a 50/50 chance of doing well. Not the best odds in every day life, but pretty good when it comes to investments (which is what schools view students as).

My recommendation to you is to trust your personal experience. If you were able to do well in high school and college, those are strong indicators that you will do well in graduate school whether or not you do well on the MAT.