Why Can't You Score a Perfect 60 on Quant or Verbal? My Theory.

Hi mcelroytutoring and HanoiGMATtutor

Hi guys, I know this thread is old, but I thought I'd throw a data point your way: I got a 51V in 2016. I later ordered my ESR and found out that I'd gotten 100% of the questions correct. So as far as I can tell, 51V must be the absolute maximum.

I was surprised to see that the difficulty level of my questions was rated as medium up until the final fourth of the exam, when it went to medium-high. How weird!

I excerpted the two pages of the ESR that deal with verbal and have uploaded them in case there's anything else there useful for you guys.

FYI, I'm a tutor, too. I've been coaching GMAT verbal since 2013, and added quant when I moved from Paris to Berlin in 2015. I've been working for other people all this time, but will soon be heading out on my own, so I'll likely become a more frequent visitor to these here parts.

Thanks for the contribution, jennpt !

Yes, ESR research so far has confirmed that you need to answer all 30 counted questions correctly on Verbal to earn a perfect V51. In fact, just 1 wrong will drop you down to a V48.

As for Quant, you can get 0, 1, or even 2 wrong in some cases and still earn a Q51. On the other hand, I have also seen an example of a Q50 with only 1 counted question wrong.

Finally, it is also possible to score Q51, V51, 790, thus missing a perfect composite score of 800 by the narrowest of margins. This occurs when you receive a perfect score on Quant and Verbal, but answer at least one question incorrectly on Quant.

In other words, you have to answer all 58 (counted) Quant and Verbal questions correctly to earn a perfect 800.

-Brian

Hi,

It's really interesting that why GMAC doesn't disclose its grading method clearly. Why it should be a secret?:) I have another question though, what do you mean by counted question?Are there uncounted questions too?

Hi Mehemmed,

Yes, there are uncounted questions--quite a lot of them, in fact (23 of 90 total).

GMAC calls them "pretest" questions, but I like to refer to them as experimental questions. They (supposedly) do not count at all toward your score and are used for research purposes only.

On the real GMAT, there are 11 experimentals on Verbal, 9 on Quant, and 3 IR, for 25.6% of the test.

-Brian

Thanks for the info. I knew that there are experimental questions on GRE, but I wasn't aware of that is the case on GMAT as well.

One of my students answered every Quant question correctly and received a Quant score of 51.
Here's the Quant portion of his ESR:

The 0-60 scoring scale is just a legacy from the old days of the GMAT. I've seen GMAT Quant scores from around 1980 (if I remember correctly) that were outside of the current 6-51 range. As I understand (it's actually hard to find information about what the test was like forty years ago) the test used to be longer, which meant that the test was able to distinguish between test takers at a greater number of levels. A short test can't tell the difference between a V51 and a V60, so there's no reason to extend the scale beyond V51. That should make sense; we know that even one wrong answer means a test taker is below a V51 level. So when a test taker answers everything correctly, the test only knows that the test taker is at least a V51. The test would have no information at all to then decide whether to give a test taker a V51, a V53, a V57 or a V60. And it wouldn't make sense to give the test taker with one wrong answer a V49, and the test taker with no wrong answers a V60, because then the test would make a minuscule difference in performance appear to be a massive difference in ability. That's even more the case when the current V51 test taker is often one who got down to 50-50 on a question or two, and made the right guesses, while the V47 or V49 test taker did the same and made the wrong guesses. To give those two test takers Verbal scores that differ by 11+ points would not correctly report the difference in their ability.

In Quant, the situation is a bit different, because Quant questions can be harder than Verbal questions, and even a short adaptive test full of hard questions can distinguish moderately well among test takers at the higher end of the scoring scale. So in theory, the Quant scoring scale probably could extend beyond the current Q51 maximum (that's clear anyway because you can get a Q51 with a wrong answer or two), but I don't see how they could meaningfully extend the scale all the way to Q60 without making the test longer, which is something they don't want to do. It's also unclear what advantage there would be, when you consider what the test is for.

I know this is an old post, but the actual explanation is much simpler.

The following information, (1), (2), and (3), is from notes I took during a 2011 meeting in which the chief psychometrician and other GMAC management team members interacted with industry professionals (instructors, consultants, and prep company owners/management). This was a relatively small discussion, but I'm putting it down here just in case any forum member (I'd be surprised if it is more than one ) is interested in this issue.

1. The scale is in fact 0-60. GMAC is not lying to all of us.
2. The actual operational portion of the scale is in fact 6-51, so the observation that both quant and verbal scores max out at 51 is also correct.

Both these points may not seem to be compatible, but all that's happening here is that the operational scale leaves a little bit of room on both sides just in case GMAC needs to extend the scale in the future. As an organisation, GMAC can't create a scale that's 0-60 and then go around telling people that it's actually 6-51, even if they're using only part of the scale. They could have communicated this more clearly, though.

There are a few additional points here:

3. They didn't think that they'd need to extend the scale any time soon. I don't know whether their position has changed, but for what it's worth, I don't think it has. As IanStewart points out, there's no real reason for them to extend the scale. The GMAT isn't designed to measure ability deep into the 99th percentile.
4. They released a document (Demystifying the GMAT: Scale Scores) explaining the scale, but I don't think they took any other action.
5. After a meeting in mid-2016 they finally changed the percentile tables on mba.com back to 6-51 instead of 0-6 and 51-60, which they had been doing 2012-2015.
6. If I remember correctly, the OGs were updated in 2018 (that'd be OG 2019).