# Tactics and Guessing

Good news. There is a strategy for guessing.

Just Mark “C”?

We’ve all been there. Sitting in the test center seemingly paralyzed as we watch the seconds tick away and realize we don’t have enough time to finish the test. Conventional wisdom has always told us to guess. “Mark ‘C,’” they say—it’s as good a choice as any. After all, the chance of getting it right is better than getting it wrong by default.

Or is it?

The sponsors of many standardized tests caught on to the game years ago. “Formula scoring” was touted as a correction for guessing. If you leave the question blank, you receive no credit for the item, but if you answer the question and get it wrong, you get no credit and you are penalized a quarter of a point (for a five-option multiple choice question—one-third of a point if there are four options). Random guessing then becomes ill-advised, and the test prep mantra became: try to eliminate the obvious wrong options before you guess to improve your odds. If you don’t have time to even read the questions, you cannot improve your odds by eliminating options. Chances are your score will be no better for random guessing—your score could be even worse.

What about the computer adaptive GMAT exam? Should you guess? Formula scoring and the correction for guessing were designed for tests where all the items carry the same weight and where you are allowed to skip questions. Throw computerized adaptive testing into the mix and you have an altogether different picture.

On the one hand, you might think that once you get near the end of the section, the algorithm has already got a pretty good handle on what your score range should be, so a question here or there shouldn’t matter. On the other hand, if you are randomly guessing near the end, you may get a question wrong that should have been easy for someone at your score level, sending the algorithm into a tizzy wondering where it went wrong and adjusting your score accordingly. The reality is, because everyone gets a different set of items, random guesses at the end will have different effects on the score. And because no one knows what item you would have gotten next if you don’t complete the items at the end, then there is no good way to estimate how that would affect your score.

Now you’re thinking: ‘Why did I read this if you’re just going to say you don’t know what would happen?’ Well, I don’t know exactly what would happen for you, but I can give you some guidelines based on our research.

First, for the obvious: Don’t run out of time! (Sorry, it had to be said.) The best defense is a good offense. Keep track of your time and you don’t have to worry about guessing or not finishing. But if you do end up pressed for time, the findings from our study may help you decide what to do.

Based on an analysis of thousands of actual GMAT records, the question of whether to guess or leave questions blank (at the end) depends on the number of items you have left, the section you are on (Verbal or Quantitative), and your relative ability. Here is how it breaks down:

• If you only have 1 or 2 items left in either section, it doesn’t make much difference if you guess or omit the question. You should finish the item you are on to the best of your ability and not worry about the others.
• If you are on the Verbal section, it doesn’t make much difference if you guess when you have up to about 5 questions left. You should finish the item you are on to the best of your ability and not worry about the others.
• In the Quantitative section, your odds improve if you guess and complete all the questions rather leave the final questions unanswered. After all, there are fewer questions in this section, so each item left blank in this section comprises a higher proportion of the test than in the verbal section. Guess as smartly as you can, but guess nonetheless—do not leave items blank.
• If you have an idea what your relative ability is ahead of time (i.e., you’ve taken a practice test or diagnostic test), then your guess versus omit strategy differs based on where you think you would fall. If your scores tend to be relatively low on the section, leaving the questions blank may actually result in a higher score than getting even the easy questions wrong by guessing. If you are near the top of the scale, you have farther to fall if you omit the items and therefore you should guess. Low ability—omit; high ability—guess; medium ability—see above.

Want to know more about the research behind these results? Read the details of the study on gmac.com.

— Eileen Talento-Miller, Ph.D., Psychometrician