Some fast facts about how the GMAT's new Integrated Reasoning section will be scored.
Fact: Right now, the GMAT has a Verbal Section (75 min), a Quantitative Section (75 min), and two Analytical Writing Assessment (AWA) essays (Analysis of Argument and Analysis of Issue, 30 minutes each).
Fact: Right now, your GMAT score report tells you: (a) V score, (b) Q score, (c) a Total score (combination of your V & Q scores), and (d) AWA score.
Fact: The "next generation" GMAT will debut on Tuesday, June 5, 2012. This test will have a V Section, a Q Section, and a single AWA essay, and the new Integrated Reasoning (IR) section. The sequence of the new test will be
1) AWA essay = Analysis of Argument, 30 minutes
2) IR section = 12 questions, 30 minutes
3) optional break, up to 5 minutes
4) Q section = 37 questions, 75 minutes
5) optional break, up to 5 minutes
6) V section = 41 questions, 75 minutes
Fact: the IR section consists of four question types
a) Graphics Interpretation (GI)
b) Two-Part Analysis (2PA)
c) Table Analysis (TA)
d) Multi-Source Reasoning (MSR)
Fact: all four question types will appear on everyone's IR sections.
Fact: the breakdown by question type will differ from one person's IR section to another person's only because of the experimental questions.
In other words, for the questions on which you are actually graded, which actually count toward your score, everyone will have the same breakdown by question type; extra experimental questions added to that baseline will give different people's IR sections different breakdowns.
GMAC has revealed neither what that fundamental breakdown is, nor how many of the 12 questions will be experimental. Let's just take a pretend scenario, just to understand. Let's say: the graded IR questions consist of 2 GIs, 2 2PAs, 2 TAs, and 2 MSRs, for a total of eight (these are my made-up numbers). For everyone taking the test, let's say those are the eight questions that are graded. The other four questions would be experimental questions, and will be different for different users. Thus, Abe might get an IR section with 3 GIs, 3 2PAs, 3 TAs, and 3 MSRs. Betsy might get an IR section with 2 GIs, 3 2PAs, 3 TAs, and 4 MSRs. Cathy might get an IR section with 2 GIs, 6 2PAs, 2 TAs, and 2 MSRs. In each case, only the baseline eight questions count toward the score, and the others are experiments. (The numbers in this example are purely hypothetical: we have no idea what GMAC has up their sleeve.)
Here's the kicker, though. As our hypothetical friend Cathy is working through her IR section, she may start to think: Gee, I'm seeing a lot of 2PA questions! Some of them must be experimental! Quite true. The catch is, among those six 2 PA questions, the two that count could be the first two, or the last two, or any combination. There are actually 15 different ways that the two that count could be scrambled among the four experimental questions. As the test taker, even if you do have strong suspicions about which question types the experimental questions were, you will have no way of knowing, as you are working on a particular question, whether it counts or is experimental. Therefore, you have to treat every single question as if it counts, same as on the Q & V sections.
Fact: the IR section is not computer adaptive. You are randomly assigned 12 questions as a group, and move through that sequence regardless of whether you are getting questions right or wrong.
Fact: The "next generation" GMAT score report will consist of (a) V score, (b) Q score, (c) a Total score (combination of your V & Q scores), (d) AWA score, and (e) IR score.
Fact: the IR score will be an integer from 1 to 8. THERE IS NO PARTIAL CREDIT ON THE INTEGRATED REASONING SECTION. There is no partial credit on the IR section. For example, in a TA question in which there are three dichotomous prompts (e.g. true/false), you must get all three right to get credit for that one question. If you get at least one of the three wrong, the whole question is wrong.
Fact: The number of IR questions you get right will constitute a raw score. The GMAC, using an arcane alchemy known only to them, will convert that raw score to a scaled score (1 – 8), which will be accompanied by percentiles.
Notice: Because of the statistical magic GMAC uses in converting raw scores to scaled scores (on IR, Q, & V sections), what may seem to your advantage or disadvantage may not work out that way. For example, the fact that there's no partial credit is challenging: it makes it harder to earn points on individual questions. BUT, harder for everyone means lower raw scores are needed to get a high percentile grade. Similarly, if all the questions are very easy, that means most people will get them right, which means it will be "crowded" at the top, much harder to place in a high percentile. What matters is not how inherently easy or hard the test is: what matters is how well you perform, compared to other test takers.
This post was written by Mike McGarry, GMAT expert at Magoosh, and is an excerpt from Magoosh's free Integrated Reasoning eBook.
Compare b-school rankings from the FT, BusinessWeek and Economist on the MBA rankings table