Brian Galvin, Veritas Prep's Director of Academic Programs, wrote this great article to help test takers understand the sometimes misunderstood GMAT scoring engine.Knowing specifically how the GMAT is scored won’t help your score much (if at all), so if you only have a few hours left to study please click to a more academic post immediately! Really the only way to improve your score is to get more questions right, so make that your primary goal as you study for the GMAT.
But discussing the scoring algorithm sure is fun, so as a quick study break that at least makes the test seem more familiar and accessible, definitely read on! The scoring algorithm is self-correcting as it narrows in on an estimate of your ability level, so it does not pay to try to game the system (by, say, spending too much time up front to ensure that you get the first 10 right -– the system will “find you out” later as you come back down to earth, and you’ll probably struggle to even finish on time in that scenario).
In this post, we’d like to clarify the curiosity of how the scaled scores turn into your overall score. Often students wonder about the scoring system when they see the same scores (e.g., two people who each score 49 Quant and 35 Verbal) correspond to two different overall scores (e.g., those same two people earning 690 and 700), or they improve their scaled scores and don’t see a significant increase in their overall (e.g., improving from 50Q and 31V to 49Q and 35V, but the increase is only from 690 to 700). What’s going on?
Dr. Larry Rudner of GMAC recently addressed these issues at last month’s GMAT Summit for top test preparation companies. Here’s what we learned from his presentation:
- The 200-800 score is the final report, sent to schools to stay consistent with the old paper-and-pencil (pre-1997) test scoring system, but the scoring algorithm used in the computer uses its own system and just converts the 0-60 and 200-800 values over as that last step in reporting.
- Each scaled score has a corresponding “interim range” (or, actually, the interim range is what the computer first gives you before converting to a scaled score) which spans 2-3 numerical values, so your 50 on Quant might correspond to a range of 60-62 on that GMAT computer scale (again, the computer uses its own numbers that are more consistent with how the CAT algorithm scores you, then the system converts to the familiar display numbers). Those interim scores for Quant and verbal are then combined and your “interim sum” (which now may span 5-6 potential values since you’re adding two ranges together) is then converted to your 200-800 score.
- So not all 49s (scaled score) are created equal — your 49 might be a “high 49″ (at the higher end of the interim range of scores and someone else’s might be a “low 49.” When the computer uses its interim value, if you have a high/high split of interim values for the same Q/V scaled scores and someone else has a low/low split on theirs, your 49/39 might be a 730 and theirs a 710. The 3-digit 200-800 score is valid — it’s just that you don’t see the full process that goes into calculating it.
So as an example (courtesy Dr. Rudner’s presentation), if you have a 50Q/31V, that has a potential range of:
– Quant: 51 scaled –> 60-62 interim range (you could be a 60 or a 62 and still end up at 51 scaled)
– Verbal: 31 scaled –> 42-43 interim range
Total: 102 – 105 on the interim scale, on which 102 would correspond to a 670 and 105 would correspond to a 690. So the same two scaled scores could swing 20 points either way on the overall score, and that’s because the overall score is calculated using the computer’s numbers, not your score report numbers.
Now, what’s most important for you to note is that the current GMAT scoring system is significantly more reliable than the old, paper-and-pencil system on which the 0-60 scaled scores and 200-800 overall scores were based. The calculation might seem convoluted, but ultimately the system’s ability to use computer adaptive testing to estimate your overall ability level provides a metric highly correlated with MBA success and highly accurate to your performance.
And, however complicated, the system is quite transparent; you can learn more about it at
https://www.mba.com, the Graduate Management Admissions Council’s site for statistical research and trends. You’ll see there that your score is calculated fairly, accurately, and efficiently. So, as always, the lesson is to worry less about how your score is measured and more about how to maximize the value of that measurement!