On June 5, 2012, the GMAT launched a new question type, the Integrated Reasoning (IR) question. The question format is highly dependent on technology, and unlike traditional Verbal and Quantitative questions, it would be too compromised in a print format. Therefore, instead of putting some compromised practice questions into the OG13, GMAC decided to launch a new website with 50 IR practice questions. When you buy your brand new copy of the OG13, that individual copy will come with an individual access code, which allows you to enter the main area of that website and practice those 50 online questions.
When GMAC first announced the IR question, it released a few free samples on its own website, but the format was not finalized at that point. Apparently, the 50 IR practice questions on this new website represents the finalized formats of each of the four IR question types, and the level of difficulty of these questions is comparable to what will appear now on IR sections on live GMATs. This post is a summary of what we know about the IR question formats, including what we can glean from these 50 practice questions.
Question Types and Terms
Each IR question is one of four formats: (1) MultiSource Reasoning (MSR); (2) Table Analysis (TA); (3) Graphics Interpretation (GI); and (4) TwoPart Analysis (2PA). If these are unfamiliar to you, you can read about the basics in the Magoosh IR eBook. On the new website, questions #118 are MSR, questions #1924 are TA, questions #2534 are GI, and questions #3550 are 2PA. On this website, they are neatly sorted by question type: we have no idea whether those questions types will be similarly sorted or completely interspersed on the IR section of a live GMAT. We also have no way to guess the relative proportions of each question type to expect on the live GMAT.
We do know: eight of the 12 questions will count, and the other four will be experimental questions. We do know: whatever block of 8 questions counts, those 8 questions will have the same distributions of question for all test takers. We don’t know what that distribution is: let’s say, just for simplicity, that it’s 2 MSR questions, 2 TA questions, 2 GI questions, and 2 2PA questions. That mix will count for everyone. Now, the four experimental questions don’t count, so they could be any mix of problems: it could be one of each type, or two of one type and two of another, or all four of a single type. Suppose Fred happens to get an IR section on which all four of the experimental questions are 2PA. In the course of the 12 questions, Fred would see six 2PA questions: if he were counting, he would realize that the experimental questions would have to be 2PA, but think about it. He doesn’t start to realize something is unusually until he gets to his 4th or 5th or even 6th 2PA. At that point, he knows some of those 2PA questions had to be experimental, but he has no way of knowing which two count: the first two? the last two? the third and the fifth? Of course, of six questions, there are 6C2 = 15 different ways the two that count could be distributed among them, and all 15 of those scenarios are equally likely. The upshot is: even if you have an inkling that you are getting more of this question than that, you always have to treat the question in front of you as if it counts.
Also, let’s talk a little about this word “question” — we say the website has 50 practice “questions” and the IR section has 12 “questions”, but I would argue a more accurate term would be “screens.” The website has 50 screens, and the IR section of the GMAT has 12 screens. Each screen will be one of four formats — MSR, TA, GI, 2PA — and almost every screen has more than one question on it. Rather than talk about how many questions within each question, for clarity, for the remainder of this post, I will refer to each “screen” and how many questions on that screen. Here’s a little of what we can glean about the four screen formats from the 50 IR practice screens on the website.
MultiSource Reasoning
This is the only of the four types in which the same content appears across multiple screens. For every other question format, all the relevant content appears on one screen, and none of that content appears on any other screen. Some MSR screens have a single fivechoice multiple choice (MC) question: on the whole IR section, this is the only screen type, the only question type, on which there is only one question on the screen. The others, in fact, the majority of MSR screens, have MultipleDichotomous Choice (MDC) questions: each MDC screen had three individual MDC questions.
On the 18 screens devoted to MSR, there were four different prompts, and multiple screens for each prompt. Here’s the breakdown:
Prompt

screens

# MDC

#MC

hotels for a conference

#16

4

2

the Nairobi Stock Exchange

#79

2

1

sports & broadcasting rights

#1012

2

1

WHO study of the growth of children

#1318

4

2

As we can see, the ratio of MDC to MC is 2:1, which makes me strongly suspect that something like this will be the ratio on the live GMAT: certainly, expect more MDC than MC. A bank of six screens on a single topic might be a little excessive: after all, that would be half the IR section right there. I would think of this like the long RC passage in the OG that have, say 6 questions: no RC passage on the live GMAT will have 6 questions, but we get the extra questions in the OG for additional focused practice. Similarly, here, with extra MSR screens.
All MSR prompts had two cards: none had three cards, so apparently this is the standard. The first and second prompts each involve one card with text and one with a table. The third prompt has nothing but text on both cards. The fourth prompt has text + a table on the first card and text + a graph on the second card. This is the “integrated” part of IR: integrating mathematical information with verbal information. While an “alltext” MSR prompt (like the third one here) may appear, apparently the majority of MSR prompts will involve tables or graphs as well as text.
Table Analysis
Few surprises here. Each TA screen contains a verbal prompt and a sortable table. Most of the tables have 78 columns, but one had 5 columns and one had 6 columns. On each TA screen, there are 3 MDC questions.
Graphic Interpretation
Each GI screen presents a verbal prompt and a graph. The questions on a GI screen are in the form of two dropdown menus in a fillintheblank format. That is to say, underneath the graph will be one or two sentences, with a total of two blanks: the student “fills in” the blank with a choice from the dropdown menu: most of the dropdown menus have 4 choices, although some had 3 choices, and a few had 5 choices. Here are the graphs that appeared:
#25 = a column chart (i.e. a histogram)
#26 = a pie chart
#27 = doublesided* scatterplot
#28 = a clustered column chart
#29 = scatterplot with two different populations
#30 = three timeplots on the same graph
#31 = doublesided* column chart
#32 = segmented column charts
#33 = four timeplots on the same graph
#34 = a clustered column chart
For more on column charts, clustered column charts, and segmented column charts, see this post. It’s reasonably clear that they love their column charts!
By “doublesided”, what mean is: there are two yaxis scales, one on the left side of the graph, and another on the right side of the graph. For example on screen #27, there are orange dots and black squares on the scatterplot: the orange dots are read on the left yaxis scale, and the black squares are read on the right yaxis scale. It’s a slick way of showing the interrelationship of three different variables.
TwoPart Analysis
In some ways, 2PA is the most flexible of the four question formats. Individual 2PA screens can be purely verbal (similar to RC and CR questions) or mathematical. This table summarizes these 16 screens according to type and number of answer choices (#A), with some comment on the nature of the “two parts” of the question. Here are my codes for the 2PA screen types:
V = completely verbal
MA = mathematical, with algebraic/variable answer choices
MN = mathematical, with numerical answer choices
Screen

Type

#A

Comments

#35

V

6

one true, one false 
#36

MN

5

scenario links the numbers needed 
#37

V

5

one cause, one effect 
#38

MN

5

scenario links the numbers needed 
#39

V

6

assumption required, possible fact 
#40

MN

6

largely a verbal question, describing a corporate hierarchy, and the question is about how many people would be involved in different review processes 
#41

V

5

a characteristic of pottery, and a prediction possible about group that made the pottery 
#42

MA

6

a geometry problem 
#43

V

6

two academics disagree: the two parts play into this disagreement. 
#44

MN

5

a distanceratetime problem 
#45

V

5

one cause, one effect 
#46

MN

5

two terms of a recursive series 
#47

V

5

most strengthen, most weaken 
#48

V

5

penalties for missed work time, which merits a verbal reprimand, which merits a written reprimand 
#49

V

5

a ballet company, what will decrease expenses? increase audience size? 
#50

V

5

advertising in two cities, what should be done in each city to test a hypothesis? 
Of these 16 questions, 10 are verbal and 6 are mathematical in some way. I would take that as an indication of the relative proportions of verbal 2PA screens to mathematical 2PA screens. All have either 5 or 6 choices, and the skills required vary from RC & CR skills on the verbal types to assorted Quantitative skills on the mathematical type.
Summary
This analysis of the 50 practice IR questions on gives us a much more precise idea of what to expect on the real GMAT. Read the Magoosh IR eBook, study this post, work through the 50 questions on the GMAC IR website, and practice the Magoosh IR questions — if you do all that, you will be wellprepared for your 12 IR screens on test day.
This post was written by Mike McGarry, GMAT expert at Magoosh and originally posted here.
[0] Comments to this Article