GMAT Focus Percentiles vs GMAT Classic

Question

Would anyone smart be able to tell me how GMAT Focus is planning to make a huge PIVOT in Percentiles while using the exactly same questions on the test?

I have overlaid percentiles for the Quant section: GMAT Focus with the GMAT Classic percentiles and ended up with this interesting chart for Quant. I am just trying to understand how they were able to chop off the top and spread it so nicely? Any ideas? To me the new percentiles are either unreliable or the scoring is really going to be extremely different... I have been pondering about this for a week now

Note : the scale for GMAT Classic (6-51) was condensed to fit into the new more compact scale of GMAT Focus (60-90) focs-vs-classis-percentiles.gif Verbal Percentiles: https://gmatclub.com/forum/download/file.php?id=72997 I found it interesting that on VERBAL, the percentiles are being bunched together. Look at the distribution of the Standard Deviation interval (hopefully calculated correctly) Classic : V20.5 - 37.5, that's 39% of the total scale (6 - 51) Focus : V74.3 - 83.3, that's 32% of the total scale... maybe not such a big deal ( Lots of people will end up with Verbal Scores between 78 and 82. I feel this is due to the score being more driven by mistakes rather than the position of the mistakes in the test (though mistakes in the test should still have an impact as the test is adaptive but just not as profound)

NaveenGJ · Answer

I have plotted scattered plot of the classic vs GMAT FE. The scatter plot for the Quant section of the Old GMAT and the GMAT FE do reveal some noteworthy differences that can provide insights into the "toughness" level of the tests and how they distinguish between test-takers. Here are some key points to consider:

Scoring Mechanics and Distinguishing Toppers:
Old GMAT : The percentile jumps significantly within a narrow score range (especially around scores 48-51). This implies that the Old GMAT Quant section may not effectively distinguish between toppers, as many end up clustered around the same percentile. It might also suggest that scoring incrementally higher becomes increasingly challenging.

GMAT FE : The percentile appears to increase more linearly with the Quant score, which suggests a more uniform distribution of test-takers across percentiles. This makes it easier to differentiate between high scorers, as the percentile more accurately reflects the score.

Insights into Toughness:
Uniformity in GMAT FE: The more linear and uniform percentile distribution in the GMAT FE suggests that the test might be designed to be more balanced across its scoring range. This could mean that it's less "tough" to move from a low percentile to a moderately high percentile.

High-Scoring Zone: Given that the Old GMAT had a clustering effect at the top percentiles, it might have been more difficult to break into the very top percentiles due to the finer granularity of scores within that range. On the other hand, the GMAT FE, with its more linear percentile distribution, may offer a clearer path to achieving a high percentile, provided one can secure the corresponding high score.

Ceiling Effect: The Old GMAT exhibits a strong ceiling effect, where a slight increase in score could result in a large percentile jump. This could make it "tougher" to score high as you would need to clear a higher bar to distinguish yourself. The GMAT FE seems to mitigate this issue to some extent.

Differential Scoring Impact: In the Old GMAT, given the tight clustering of top percentiles, even a one-point increase could have a disproportionately high impact. In the GMAT FE, a more balanced scoring system might mean that each point increase has a more uniform impact on your percentile ranking.

Final Thoughts:
The GMAT FE seems to provide a more balanced and differentiated scoring system, potentially making it "easier" to distinguish yourself if you are a high scorer. However, the term "toughness" could be relative; the GMAT FE might be less forgiving of mistakes for those aiming for the very top percentiles, given the more linear percentile-to-score relationship.

bb · Answer

Thank you  NaveenGJ , interesting first post. Is this a copy paste from somewhere?

NaveenGJ · Answer

No it isn't a copy paste. I wanted to post the graph or its imgur link as well but I can't as I haven't posted 5 comments on the forum.

Have done similar analysis of Verbal and overall GMAT FE vs Classic GMAT as well.

Posted from my mobile device

bb · Answer

Awesome! Thank you for a very thoughtful reply and kudos for doing an analysis!

PS. What does the verbal chart tell you?

NaveenGJ · Answer

Old GMAT:
Concentration at the Top: In the Old GMAT, Verbal scores ranging from 45 to 51 are all in the 99th percentile. This concentration makes it challenging to distinguish among top performers who are all grouped into the same percentile. This lack of differentiation makes it less rewarding for incremental improvements at the high-end.

GMAT FE:
Expanded High-Scoring Range: In the GMAT FE, the 99th percentile starts at a Verbal score of 87 and extends to 90 for the 100th percentile. This expanded range allows for finer granularity among top scorers.

Clearer Incremental Progression: For instance, a one-point improvement from 86 (which is in the 98th percentile) to 87 moves the test-taker to the 99th percentile. This more clearly rewards incremental improvements.

I would have loved to post a graph for elucidating above points, but I can't because of the '5 post policy'.

NaveenGJ · Answer

Here are the charts showing comparison between Classic and Focus versions of GMAT score & percentile for quant, verbal and overall scores.

bb · Answer

Thank you for doing this!!! I like your graph style. Very clean!

Here is my Verbal chart with some finagling of scales.

Attachment:

verbal-focus-classic-percentiles.gif [ 46.81 KiB | Viewed 9480 times ]

(Note that Classic X axis is at the top of the chart and the Focus X-axis is at the bottom)

I found it interesting that on VERBAL, the percentiles are being bunched together. Look at the distribution of the Standard Deviation interval (hopefully calculated correctly)

Classic: V20.5 - 37.5, that's 39% of the total scale (6 - 51)
Focus: V74.3 - 83.3, that's 32% of the total scale... maybe not such a big deal (

Lots of people will end up with Verbal Scores between 78 and 82. I feel this is due to the score being more driven by mistakes rather than the position of the mistakes in the test (though mistakes in the test should still have an impact as the test is adaptive but just not as profound)

Attachment:

histograms.gif [ 65.63 KiB | Viewed 7890 times ]

At the same time, Quantitative scale will see the opposite effect - QUANT percentiles are being spread out.
So the same change has opposite effects on the Quant and Verbal sections.

So at the end, the result is the opposite, but largely unchanged

Candidates will have a bunched up Verbal Score and a Diverse Quant score instead of previously having a bunched up Quant.

On the positive side, I think it is going to be great to see some more space available for people to improve their Quant rather than being stuck with Q51 or Q50. (it was previously possible to get Q51 with 1-2 wrong but I guess now that will be a more stretched out space)

Candidates will have a bunched up Verbal Score and a Diverse Quant score instead of previously having a bunched up Quant. On the positive side, I think it is going to be great to see some more space available for people to improve their Quant rather than being stuck with Q51 or Q50. (it was previously possible to get Q51 with 1-2 wrong but I guess now that will be a more stretched out space)

MartyMurray · Answer

I think the Quant section will achieve the new percentiles by just dinging your score for each missed question, scoring differently from the classic GMAT, on which you can score Q49 with 10 missed Quant questions, Q50 with a few missed questions, and even Q51 with a missed question or two.

On the Focus Edition, missing one Quant question will knock your score down to the upper 80s, missing two will knock it down to the mid 80s, etc. So, scores will be distributed more widely. Only test-takers who get all but one or two questions correct will get top scores. There won't be people missing eight questions and scoring within two points of the highest possible Quant score.

raju1212 · Answer

But don't you think things will be more dynamic, since this test is also sectional adaptive. So a person starting with VA DI and scoring very high in them 85+, will get relatively tougher question in QA, and perhaps can afford to make more mistakes?

Sim_SG3 · Answer

Wait, I thought the adaptive nature of the test wasn't across sections but just within the section? If the 3 sections are not treated separately, especially in terms of the questions being adaptive, there's no point in giving each section equal weightage in determining the overall score.

Someone doing exceedingly well in the first two sections would face hell in the third round and would be more likely to mess up. It'd be cruel to make the hard section count just as much as the two good ones.
Conversely, doing poorly in the first two sections would mean you'd only get low-level questions in the third section and would have virtually no way of saving your score.
Imagine equal weightage on top of all of that!

Sim_SG3 · Answer

Closing the loop on this - the exam isn't adaptive across sections. Hurray! Source: https://www.youtube.com/watch?v=lbb3Rw5SOzA Thanks to BottomJee for sharing the above and putting an end to my misery. Best, Sim

raju1212 · Answer

https://youtu.be/Hfgby9-M__o?t=2746
Check this video.. it is in a way section adaptive. but the impact wont be significant. With adaptive questions getting harder, i just meant that there might be more room to make mistakes if average difficulty goes up, so we cant have pin point surety like in the classic,of how many mistake one can afford to make.. .

AjiteshArun · Answer

Hi raju1212,

Both the GMAT and GMAT Focus are question-adaptive. Here are some points we should keep in mind:

1.  "Section-adaptive" means that  the entire next section is selected  on the basis of the test taker's performance in an earlier section or sections. This is not what the GMAT does. The GMAT incorporates information from the previous section in the selection of only the first question of the next section.

2.  Looking at the previous section as a whole can be misleading. For example, some people think that the GMAT chooses a question on the basis of the test taker's performance on the previous question. This is incorrect. Both the GMAT and the GMAT Focus Edition use the test taker's performance on  all the questions up to that point  in that section to choose the next question (there are some additional constraints). So, to a certain extent, what the Focus Edition is really doing is that it's not treating Q/V/DI as being completely separate.

You are absolutely spot-on about accuracy not being a reliable predictor of score on an adaptive test like the GMAT/GMAT Focus.

Sim_SG3 · Answer

Ah, I understand now. I took "section adaptive" to mean what  AjiteshArun  described in his reply to this post. Thank you for clarifying!

Best,
Sim

	Magoosh increases your score, guaranteed - 15% off ALL Magoosh GMAT plans with code GMATCLUB15 + Free GMAT Club Tests.
	Manhattan Prep: $79 Bonus Score Higher on Your GMAT Exam. Guaranteed.
	Target Test Prep - Save up to $325 $200 Discount + $79 Bonus GMAT Club Tests
	e-GMAT: 60% Off on GMAT Focus Get to 735+ from any starting point
	Experts' Global - ~~$1720~~ $960 End-to-End GMAT Prep + App Support: 150+ Videos, 2000+ Questions, 15 FLTs +
	GMATWhiz - $299 ~~$499~~ Experience the course through a free trial.
	GMAT Club Tests - Free Included with every course purchase of $149+ View all deals

GMAT Focus Percentiles vs GMAT Classic

Prep Toolkit

Top 5 GMAT Debrief Videos

615 to 715 in 15 Days (Ria's Strategies)

More than Quant/Verbal Abilities: Speed vs. Accuracy

745 with Only Self-Prep and GMAT Club

From 555 to 765: 210-point Improvement

Perfect 805 Score Debrief by Julia

My Rewards