Tag Archives: GRE percentiles

GRE vs. GMAT, part VI: what is a good GRE score, anyway?

 

In the fifth installment of my (apparently never-ending?) GRE vs. GMAT blog series, I suggested that the GRE might offer an advantage if your GMAT score is on the low side, since the GRE doesn’t yet appear in the MBA rankings. So now you might be wondering: what sort of GRE score could be considered “good enough” for top MBA programs?

Unfortunately, MBA admissions committees aren’t exactly open about their handling of the GRE. The good folks at Poets & Quants recently published a few trickles of GRE data from top MBA programs, but the data is still lamentably limited.

So how might MBA admissions committees define a “good” GRE score, and what sort of GRE score should you be aiming for? Let’s start by laying out the three main ways that MBA programs could evaluate GRE scores.

Option #1: the GMAT-GRE conversion chart

As the GRE made its push for acceptance in the MBA world, it published a handy little GMAT-GRE conversion chart, so that you can take any GRE score and turn it into an “equivalent” GMAT score. Sounds great, right?

Trouble is, conversion charts that compare two different tests are on shaky scientific ground. In theory, the GRE and GMAT are designed to measure subtly different skills, and they’re on completely different score scales. Few, if any, independent psychometricians (experts in the statistical science underneath standardized testing) would suggest that these conversion charts should ever be used. They simply aren’t very accurate, even under the best of conditions.

The GMAT-GRE conversion chart was, of course, published by the creators of the GRE, who have made an aggressive attempt to seize market share from the GMAT. You won’t be surprised to hear that the GMAT responded with an anti-conversion-chart article in an old GMAT newsletter from 2009.

More detail is available on the GMAT website, but here’s the bottom line: GRE and GMAT scores are correlated, but they aren’t perfectly correlated. If you’re trying to convert GRE scores into GMAT scores, the standard error of prediction is 67.4 points, according to GMAC. In other words: if MBA programs try to use your GRE score to predict what you would have gotten if you’d taken the GMAT instead – the conversion chart is likely to be off by an average of 67.4 points in one direction or another.

And in case you’re new to the (painful) reality of MBA admissions: 67.4 points on the GMAT is a really, really big deal these days.

So if MBA admissions committees are using the GMAT-GRE conversion chart… well, they shouldn’t be.

Option #2: GRE percentiles

GRE percentile scores would, on the surface, appear to be a fairer way to evaluate MBA applicants who have taken the GRE. Why wouldn’t MBA programs just look at applicants’ quant and verbal percentiles on the GRE, and then compare them to the equivalent GMAT percentile scores?

But there’s a huge problem here, too: percentile rankings compare you with other people who took the same test. So if, for example, MBA admissions committees started using the dreaded 80th-percentile rule for GRE scores, that would be unfair: your GRE percentile rankings would be based on the scores of the 700,000+ GRE test-takers – only a small percentage of whom are applying to MBA programs. That’s not cool: you’ll face a completely different horde of competitors for the GRE vs. the GMAT.

And of course, percentiles are out of whack on the GMAT quant section, anyway. If you want to score in the 80th percentile on the GMAT, you’ll need a 49 – which requires some serious skill. On the GRE, all you’ll need to hit the 80th percentile on quant is a 162 – and that’s nowhere near as difficult as getting a 49 on the GMAT.

In other words, GRE percentiles absolutely cannot be compared fairly to GMAT percentiles. It’s a terrible idea to do so, and I desperately hope that no MBA admissions committees have gone down this road.

Option #3: Round numbers, human nature, and the GRE 160 sniff-test

Here’s where we get into the interesting stuff. I’m sure that some MBA admissions committees are making the mistake of using the GRE-GMAT conversion chart, and others might be (mis-)using GRE percentile scores. But my hunch is that most MBA admissions committees probably just use a “sniff test” for GRE scores – and I’m guessing that a 160 on both GRE sections is probably enough to make MBA adcoms move on to other aspects of your MBA application.

Let’s face it: people are naturally drawn to nice, round numbers. For a long time, a 700 was considered a key threshold on the GMAT, partly because it’s a nice round number, but also because the creators of the GMAT were thoughtful about that round number: if 500 was supposed to be the mean score on the test, a 700 was supposed to be roughly two standard deviations above the mean – and well above the 90th percentile.

GMAT scores and percentiles have become pretty warped, of course, but the GMAT score scale was originally designed wisely: a 700 wasn’t just a nice, round number. It was also a meaningful dividing line between high scorers and ridiculously high scorers. And the same score scale was used on both the SAT and the pre-2011 version of the GRE – so the GMAT’s 200-800 scale is a familiar friend that has been part of American education culture for decades.

But our intuition goes out the window with the GRE’s new score scale, which ranges from 130 to 170 for each section. If 150 is the average GRE score, then what’s our instinctual dividing line between “very good” and “elite”? I suspect that MBA admissions officers struggle to understand the difference between, say, a 157 and a 159, but a 160 sounds nice and round, right? So I think that adcoms tend to feel OK about an applicant’s ability once they’re above the 160 mark on both the quant and verbal sections. No other implicit “cut score” would seem to make much sense.

So if you’re not much of a GMAT-slayer, my advice is to try to crack 160 on both the quant and verbal sections of the GRE. Unless you have an unusual set of strengths and weaknesses, it’s much, much easier to get a pair of 160s on the GRE than it is to get, say, a 700 on the GMAT – and of course, a 700 on the GMAT might not be enough anymore, anyway.

For whatever it’s worth, the few GRE averages that have leaked out might support the idea that MBA programs are using the GRE as a “way around” the GMAT/MBA rankings game – and it might also support the idea that a 160 could be enough to make your test score a non-issue at top-tier MBA programs.

For example, Michigan, Cornell, and UCLA all had average GRE verbal scores of 162 and average quant scores of 159; Duke’s averages were 160 on both the quant and verbal GRE sections. That’s certainly not conclusive, but it suggests that MBA admissions committees might be using 160 as a loose GRE benchmark score — even though the GMAT-GRE score conversion chart says that a pair of GRE 160s is equivalent to a not-terribly-competitive GMAT composite score of 640.

The GRE score data remains limited, so take it – and my interpretation of it – with a grain of salt. But if you’re OK with a test score that makes your MBA application “GMAT-neutral”, then the latest GRE data offers some interesting food for thought.