Myth or Fact: RC=3×SC
Is RC really more important than SC and CR?
Have you come across some advice that encourages you to work harder on, and perhaps prioritize during the actual exam, a particular question type: RC? It often takes this form: if you “have” to get X questions wrong, it’s better to make sure that those X questions are SCs or CRs. Some even go so far as to say that individual RC questions are twice as important as CR questions and thrice as important as SC!
Maybe some people ran GMATPrep a few times and noticed that they got a higher score when when their RC accuracy was better. Maybe when they did badly in RC their overall score dropped more than it did when they got the same number of SC or CR questions incorrect. Whatever the reason, they are convinced that each RC question “counts for more” on the GMAT exam. So, is RC the most important question type on the GMAT?
Now, it is quite clear that RC tests basic skills that can’t be improved overnight, and I’m not trying to get into how much effort you, as an individual, should put into getting better at RC. What we’re looking at here is whether you should prioritize RC over SC and CR during your test. Here are three things we should look at:
1. The GMAC’s position: the chief psychometrician at the GMAC has confirmed multiple times in meetings with industry professionals that the GMAT does not assign greater weight to RC. It’s the psychometric properties of a question that determine how the algorithm looks at that question. The algorithm ensures that test takers receive a balanced test, but it does not take question type into account during the scoring process.
2. The way an adaptive exam works: an adaptive exam looks at the likelihood of getting a particular response pattern at different ability levels, and then chooses the level with the highest probability.
What does that mean? Let’s say that in a particular GMAT exam only two questions are left, one an extremely difficult (“800 level!”) SC question, and the other an extremely easy RC question (ignore the fact that each passage has a “block” of questions).
Any test taker who is doing reasonably well will be penalized more for getting the RC question wrong than for getting the SC question wrong. But this is because of the difficulty level of the question, not the question type itself! We can look at this as a situation in which the exam expected the test taker to get the tough question wrong, but did not expect the test taker to get the easy question wrong. The first type of mistake would not cause the score to change, but the second type would!
3. Data from ESRs: the Enhanced Score Report is not perfect, but it does give us (a) performance across SC, CR, and RC and (b) total accuracy (total correct and therefore total incorrect). Can we find even one data point that shows that the same score can be achieved with the same number of questions incorrect and lower performance in RC?
Before we start, the highest possible percentiles are (a) CR 97-98, (b) RC 95, and (c) SC 99. The following examples are of different test takers with the same total verbal score and the same number of questions incorrect.
Here test taker 1 has got every RC question correct while making 2 mistakes in SC and CR. Test taker 2 has got every SC and CR question correct while making 2 mistakes in RC. So, the same number of mistakes, with different distribution, but the same total score.
Here also we see perfect performance in RC with 6 mistakes spread across SC and CR. Test taker 2 has the same SC percentile, but a much lower RC percentile (again with a total of 6 questions incorrect).
In this case the second test taker has lost a lot of ground in both RC and CR, but his or her final score is the same because of an excellent performance in SC. If SC were really not important, it would have been impossible for test taker 2 to bounce back.
So, what should you take away from this?
1. RC is not unimportant. It’s just not more important than the other question types.
2. Also, do not conclude that accuracy alone determines the score. That's not how an adaptive test works, and different ESRs will show different outcomes.
3. Don't sacrifice your SC and CR on test day in an attempt to boost your score.
If your SC and CR are already very good, but your RC isn’t, you should be spending more time on improving your RC. Unless you are very strong at a particular question type, work on all 3 (SC, CR, and RC). It’s a very good idea to build your skills across all question types.
Additional note: any effort you put into developing your reading skills will pay off over the long run. Reading is one of the most important skills to develop if you’re pursuing excellence, and it is one of the skills that enable you to grow as an individual and a professional.