GMAT Critical Reasoning Inference Questions

by on August 1st, 2012

What would you infer if I told you that I don’t eat chocolate ice cream?

You might guess that I’m on a diet, perhaps, or maybe that I’m lactose intolerant. Or maybe it’s not the milk that’s the problem; I could be deathly allergic to chocolate. Or, you might infer (correctly) that I just don’t like the flavor.

What could you infer if the GMAT told you that I don’t eat chocolate ice cream?

You can infer that if I eat ice cream, I will always choose a flavor other than chocolate. And that’s about it.

The Inference category of GMAT Critical Reasoning questions asks you to make logically supported inferences. You take the text of the stimulus at its word (recognize these questions by language such as “If the statements above are true”), and find the answer choice that must be true on the basis of the prompt.

In your GMAT prep, you will find that the biggest challenge to solving Inference questions is that there are lots of things that could be true. Sometimes, you can cleverly piece together a puzzle and make a solid prediction. But unlike argument-based Assumption questions, Inference Q’s don’t always lend themselves to knowing the answer before you look at the choices. For instance, if I don’t eat chocolate ice cream, you can infer that I wouldn’t eat chocolate ice cream cake (which contains chocolate ice cream), that a friend who knows my dietary preferences wouldn’t buy me a scoop of chocolate ice cream (which I wouldn’t eat), and that I am more likely to be seen eating vanilla ice cream than eating chocolate ice cream (because I may or may not eat vanilla ice cream, but I certainly will not eat chocolate ice cream). Since any of these would be an acceptable answer, but only one can appear in the answer choices, trying to pin down the right answer without looking at the choices can be inefficient. Inference questions are the only CR question type where you should plan to go through all five answer choices looking for one that sounds good.

But be aware of out-of-scope traps. You have to go by what the text tells you, and nothing else. And you must be able to determine the correct answer with certainty. In the chocolate ice cream example, you don’t know if the “chocolate” or the “ice cream” is the reason that I don’t eat chocolate ice cream (or something else entirely!). You might guess that I prefer vanilla ice cream, but maybe I can’t digest the milk in any type of ice cream. You might suppose that I don’t like chocolate, but it’s possible that chocolate only tastes bad to me in ice cream form and I’m fine with chocolate bars and chocolate chips. These are the types of reasonable suppositions you might make in real life, but not the type of Inference that the GMAT requires you to make.

Try today’s practice problem below. Stay in scope of the text, and don’t let reasonable, but uncertain, predictions of the future distract you from your goal.


A new electronic security system will only allow a single person at a time to pass through a secure door. A computer decides whether or not to unlock a secure door on the basis of visual clues, which it uses to identify people with proper clearance. The shape of the head, the shape and color of the eyes, the shape and color of the lips, and other characteristics of a person’s head and face are analyzed to determine his or her identity. Only if the person trying to open a secure door has the required clearance will the door unlock. Because this new system never fails, an unauthorized person can never enter a secure door equipped with the system. If the statements above are true, which of the following conclusions can be most properly drawn?

(A) The new system is sure to be enormously successful and revolutionize the entire security industry.
(B) The new system can differentiate between people who are seeking to open a secure door and people passing by a secure door.
(C) No two people have any facial features that are identical, for example, identical lips.
(D) High costs will not make the new security system economically unviable.
(E) The new computer system is able to identify some slight facial differences between people who look very similar, such as identical twins.


Step 1: Identify the Question Type

Since the stem asks us to accept the statements as true and draw a conclusion on the basis of them, this is an Inference question.

Step 2: Untangle the Stimulus

The stimulus tells us that a new electronic security system is completely failsafe and will never allow an unauthorized
person through a door equipped with the system. And the system allows an authorized person to enter solely on the
basis of the person’s appearance and facial features.

Step 3: Predict the Answer

Attempting to predict the correct inference could waste time, but on the GMAT, to make an inference means to
determine what must be true, not just what could or might be true. It’s crucial to approach the answer choices with
this in mind.

Step 4: Evaluate the Choices

(A) is out of scope. We have no evidence of how the security industry is going to respond to the new system.

(B) doesn’t need to be true. The new system doesn’t need to differentiate between people passing by the door and people trying to enter, as long as it lets authorized people in and keeps unauthorized people out.

(C) is too extreme. We don’t know that any one feature cannot be the same. All we know is that all of the features can’t be the same. According to the stimulus, the security system examines multiple facial features to determine identity.

(D), costs are outside the scope of this stimulus, since the stimulus only discusses the likelihood that unauthorized people will be able to get past the security system and through a secure door.

(E) If one twin is authorized and the other isn’t, we know the door must be able to tell them apart, because the stimulus tells us that the security system never fails. Thus, (E) must be true.


  • Good tips.broadens the perspective.

  • nice distributed way of facing such kind of problems. 

  • good one ! helps in identifying the conclusion in the options given.

    As a learning, any option that is remotely close to any one of the premises cannot be a correct answer. As option "B" borders on saying that same thing as mentioned in premise - sentence 2 & 3, it is out. Option "E" fits just right in this example.

    • Hi Rohan,

      Be careful with your assumptions! It's not the case that you can rule out a choice for being "too simliar" to the text. Part of what makes inference questions so hard to predict is that anything close to any one of the facts in the stem can be the basis for a logical inference!

      (B) is wrong because there is no evidence that the system can even detect people passing by the door, let alone need to distinguish those people passing from those opening the door. All we're told is that if and when someone tries to open the door, the system determines if they are authorized. 

      Hope this helps, and good luck studying!



  • This has been very comprehensively explained.thanks

  • Thanks Eli ! I stand corrected here. Looking back, I think I got jumbled up in words and did not critically analyze option B. This is a classic example of "haste makes waste".

    • Very well explained ! This is one area in CR where I lack. Thanks.

  • Isn't E is an assumption on which conclusion should be based upon ??

  • Nice one bt can u pls give another example

  • Hi! Quick question that's totally off topic. Do you know how to make your site mobile friendly?
    My weblog looks weird when browsing from my iphone4. I'm trying to find a template or plugin that might be able
    to correct this issue. If you have any recommendations,
    please share. Many thanks!

    Feel free to visit my webpage Content Management System Cost

Ask a Question or Leave a Reply

The author Kaplan GMAT gets email notifications for all questions or replies to this post.

Some HTML allowed. Keep your comments above the belt or risk having them deleted. Signup for a Gravatar to have your pictures show up by your comment.