In the late 1960’s and early 1970’s, Stanford professor Walter Mischel conducted one of the most famous social experiments of all time.   Known as the “Marshmallow Test,” the experiment worked like this:

A child was brought into a room and a marshmallow was placed in front of that child.   The experimenter told the child that he would return in 15 minutes, and if the child did not eat the marshmallow before his return, then that child would be given two marshmallows.

Of the 600 children (mostly 5 year olds) involved in the experiment, only about 1/3rd waited the 15 minutes to get the two treats.   A full 2/3rd of the children ate the marshmallow before the experimenter returned.    All of this was videotaped and as you might imagine the videos are hysterical:  some children pop the marshmallow in as soon as the experimenter leaves; many handle the marshmallow, take little bites, and then finally give in completely before the 15 minutes are up; some stoically wait the 15 minutes and happily take their two marshmallows when the experimenter returns.

This experiment on delayed gratification became famous mainly because of the surprising data collected by the experimenters after following these 600 children over the course of their lives.   The 200 children who were able to delay their gratification and get two marshmallows had vastly better life outcomes:  higher SAT scores, higher educational attainment, lower body mass index, and superiority in many other life measures.

Interesting for sure, but how does this relate to the GMAT? In watching thousands of GMAT students over the years, I have noticed that those who are too eager for gratification – those who are too quick to attack and answer problems without first deliberately considering them – have very poor GMAT outcomes.   However, those who delay their gratification – those who patiently consider the problem and all answer choices first – have very good GMAT outcomes.     The writers of GMAT questions are well aware of the “Marshmallow Effect”:  they know that students will impatiently attack questions or go immediately to statements in Data Sufficiency before they really understand what the question is asking and before they have considered the best approach.

Consider this difficult data sufficiency question:

The infinite sequence a1, a2,…,an,… is such that a1 = x, a2 = y, a3 = z, a4 = 3, and an = an-4 for  n > 4.  What is the sum of the first 98 terms of the sequence?

(1)       x = 5
(2)       y + z = 2

(A)  Statement (1) ALONE is sufficient, but statement (2) alone is not sufficient to answer the question asked;
(B)  Statement (2) ALONE is sufficient, but statement (1) alone is not sufficient to answer the question asked;
(C)  Both statements (1) and (2) TOGETHER are sufficient to answer the question asked; but NEITHER statement ALONE is sufficient.
(D) EACH statement ALONE is sufficient to answer the question asked;
(E)  Statements (1) and (2) TOGETHER are NOT sufficient to answer the question asked, and additional data specific to the problem are needed.

If you patiently analyze the question stem and leverage all that information first, then you will realize that the answer is (E).   If you don’t do that and rush to the statements, then you will almost surely pick (C) – the sucker choice on this problem.   A sequence such as an = an-4 means that the sequence repeats every four terms.   For instance if a1 = 2, a2 = 4, a3 = 1, and a4 = -2 then the sequence would go like this:   2,4,1,-2,2,4,1,-2,2,4,1,-2……    Every four terms would sum to a total of 5 and you could use that knowledge to quickly get the sum of the first 98 terms.   In the first 98 terms there are 24 complete sets of 4 terms (4 x 24 = 96) and then you would have to add the first and second term to get to 98 terms.   The sum would then be 24 x 5 (the sum of every four terms) + 2 (first term) + 4(second term) and the sum would be 126.   In this question, they want the sum of the first 98 terms but variables are used instead of the sample numbers given in the previous explanation.   Still you should behave the same and simplify this question.   Using the same logic as above, this question is really asking for the value of 24 sets of the four terms + first term + second term or

24(x + y + z +4) + x + y.   Here is your new question after careful analysis of the question stem:

What is 24(x + y + z +4) + x + y?

(1)       x = 5
(2)       y + z = 2

Now it doesn’t seem so hard.   Each statement alone is clearly insufficient to get a value, and the choice between (C) and (E) is much easier if you have patiently simplified the question stem.   Since you never get a value of y from the two statements the answer must be (E).   The value of that expression with both statements is 24(5 + 2 + 3) + 5 + y so the value is cannot be determined.   It is hard to realize this if you don’t use most of your time on this question assessing and simplifying the question stem itself.

The bottom line:  don’t eat the marshmallow right away.    Fully digest the problem and THEN consider the best approach.   In critical reasoning, fully understand the argument and try to anticipate gaps and flaws before you pollute your mind with incorrect answer choices.   In data sufficiency, don’t go to statements before you have fully leveraged everything given to you in the question stem.   Remember that answer choices in multiple-choice format and statements in Data Sufficiency usually provide hints, but they are not always your friends.  On the GMAT there is a great reward for those who delay their gratification – you will be more likely to get the problem correct.  And – like in the Marshmallow test – this will probably lead to greater life successes, whether it’s admittance to Harvard Business School or that top job at Goldman Sachs.

Plan on taking the GMAT soon? We have GMAT prep courses starting all the time. And, be sure to find us on Facebook and Google+, and follow us on Twitter!

Chris Kane is a longtime Veritas Prep instructor and recipient of the worldwide Instructor of the Year Award. Having taught thousands of students in New York City and the tri-state area, he contributes frequently to the Veritas Prep lesson materials and is the primary instructor for the popular Immersion Course in Manhattan, where he will begin another such course next month.