Most GMAT students respond to a plateau by studying more. Sadly, that doesn't always work, often because it's hard to know what you actually need to improve to break the plateau. What's key is having the right data. Let's use one of my students as an example.
Ethan got 6 of 9 CR questions correct. All three misses were on weaken questions. Easy fix, right? Just do more weaken questions! That's what most students would assume, but the data points to a different problem.
Ethan correctly identified 2 of those 3 weaken questions, deconstructed the argument correctly in all three, and was looking for the target in all three. He ended up getting all three wrong though. If Ethan had a structural issue with weaken questions specifically, I'd expect to see that he either wasn't recognizing them consistently or was having difficulty identifying what he needed to be looking for in the answers (his target), neither of which was the case here. So then why did he get these wrong?
The answer was in the timing data. On one question he spent only 28 seconds evaluating answer choices before moving on. On two others he spent over three and a half minutes and still got them wrong. Rushing into trap answers and getting stumped by the answer choices are both symptoms of the same issue: not conceptualizing his target in the right way: too vague and you fall for a trap answer; too detailed and you overwhelm your working memory and struggle to distinguish good answers from bad ones. The fact that these issues all happened in weaken questions was just a strange coincidence.
Most GMAT prep provides content and questions. Very little of it gives you the data to understand what is actually happening when you solve. That gap is why scores plateau.
Here's the full data behind that analysis:
Performance ReportEthan answered 6 of 9 questions correctly, falling one short of the passing threshold of 7. His three errors were all on Weaken the Argument questions.
Accuracy by Question
| Question | Subtopic | Difficulty | Ethan's Answer | Correct | Result |
|---|
| 100970 | Find the Assumption | Medium | A | A | Correct |
| 100867 | Weaken | Medium | D | D | Correct |
| 100837 | Weaken | Medium | A | D | Incorrect |
| 100653 | Weaken | Medium | B | A | Incorrect |
| 100725 | Weaken | Hard | E | A | Incorrect |
| 100649 | Strengthen | Hard | E | E | Correct |
| 100840 | Weaken | Hard | A | A | Correct |
| 100930 | Weaken | Hard | B | B | Correct |
| 100849 | Evaluate | Hard | B | B | Correct |
Medium: 2/4 | Hard: 4/5
Notably, Ethan performed better on hard questions than medium ones, suggesting his errors stem from two distinct issues rather than a general content gap.
Process Step AnalysisStep 1: Question Type IdentificationFast and largely accurate. One error: on 100653, he labeled it as Strengthen when it was a Weaken question. His Step 3 target correctly identified a weakening direction, suggesting he partially self-corrected, but the misclassification introduced confusion that carried through.
Step 2: Argument ExtractionExtraction times were reasonable at 18 to 41 seconds across valid data points.
Step 3: Target IdentificationConsistently fast at 2 to 67 seconds. Targets were well-formed on every question including the ones he got wrong. The breakdown is not in understanding what he is looking for. It is in evaluating the answer choices against the target.
Step 4: Answer EliminationTwo distinct patterns. On 100837, Ethan completed the entire question in 55 seconds with only 28 at Step 4, almost certainly reflecting insufficient evaluation before moving on. On 100653 and 100725, he spent well over the 2:00 benchmark and still chose the wrong answer, with 151 seconds at Step 4 on 100725 alone.
Timing Summary
Question | Total Time | vs. 2:00 Benchmark |
|---|
| 100970 | 2:00 | On pace |
| 100867 | 3:22 | Over |
| 100837 | 0:55 | Under (rushed) |
| 100653 | 3:44 | Over |
| 100725 | 4:08 | Over |
| 100930 | 3:14 | Over |
| 100849 | 1:09 | On pace |
5 of 7 questions with valid timing are outside the 2:00 window.
Key Diagnostics[*]1. Step 3 target is the breakdown point. Targets are accurate, but may not be clear and concise enough to facilitate efficient answer evaluation. The issue is execution at the answer evaluation stage.
[*]
2. Pacing is a concern and likely downstream of the Step 3 difficulty. Improving step 3 technique and further developing wrong answer pattern recognition should help pace as a byproduct.
If you would like to see what this kind of diagnostic looks like for your own performance, feel free to post your results below or send me a message directly.