Editorial: Companies worldwide have started adopting MAT (Ma

This topic has expert replies
Junior | Next Rank: 30 Posts
Posts: 10
Joined: Fri Apr 25, 2014 10:27 pm
Editorial: Companies worldwide have started adopting MAT (Management Aptitude Test), a decade-old standardized online-based aptitude entrance test administered centrally by CGI (Confederation of Global Industries) all round the year to screen in the most eligible candidates for their respective interviews. Latest statistics released by CGI show that the average error rate, a statistical figure to assess the difficulty level of a question and determined only after the number of respondents reaches at least thousand, of ten toughest questions in a sample group of MAT tests conducted within any pre-defined time interval not earlier than three years ago has never exceeded 70%. A particular question recently introduced by CGI in MAT was, however, found to have an error rate of more than 80%, attempted by approximately 5000 candidates worldwide. Hence, this question could be considered among the toughest of the questions to have appeared on the MAT. But, some respondents contend that the new question was indeed so tough need not be true as it is also quite likely that the official answer to the question itself is wrong.

Which of the following, if true, would most help to arbitrate the contention presented in the argument?

A. 95% of the respondents who got the question right scored at least 99 percentile on the MAT.

B. Statistics released by CGI have revealed that for any time interval put under consideration since the last ten years, the difference between the number of respondents per thousand answering the ten toughest questions correctly and number of those answering them wrong has varied between 200 and 650.

C. The average error rate of the ten toughest questions in any time interval, not later than three years ago, was approximately 74%.

D. As per latest records available by the CGI, most of the all-time toughest questions on MAT put to test, the error rate of which exceeded 80% but not 85%, have been certified as perfectly valid questions with correct official answers by more than one independent auditor.

E. Studies by certified psychometricians indicate that a question in any online-based aptitude entrance test cannot achieve a maximum error rate not lower than 80%, even after allowing for a maximum grace variation in error rate up to 5%.

[spoiler]OA:E[/spoiler]

A great question...Rack your brains!! :wink:
Source: — Critical Reasoning |

User avatar
GMAT Instructor
Posts: 2193
Joined: Mon Feb 22, 2010 6:30 pm
Location: Vermont and Boston, MA
Thanked: 1186 times
Followed by:512 members
GMAT Score:770

by David@VeritasPrep » Tue Sep 30, 2014 6:17 am
This is not in the format of a GMAT question. There are multiple problems with the language and the logic. This question was posted at GMAT Club and the consensus over there is that this is not a good question.

What is this answer supposed to mean?
a question in any online-based aptitude entrance test cannot achieve a maximum error rate not lower than 80%, even after allowing for a maximum grace variation in error rate up to 5%.

This does not make sense.

My advice is to ignore this question. This question is a great example of exactly what the GMAT will never do. That is try to make a question difficult by making it unnecessarily convoluted.
Veritas Prep | GMAT Instructor

Veritas Prep Reviews
Save $100 off any live Veritas Prep GMAT Course

Junior | Next Rank: 30 Posts
Posts: 10
Joined: Fri Apr 25, 2014 10:27 pm

by itzmyzone911 » Tue Sep 30, 2014 8:06 pm
David,

I do not agree that the answer does not make sense at all, although I agree that there could be a fair amount of concerns regarding the convolutedness of this question, making it a bit GMAT-unlike.

Here's my inference about option E.

x-x-x-x-x-x-x-x-x-x-x-x-x-
It states that studies by certified psychometricians (those dealing with measurement of human intellect vide IQ and aptitude tests, or those dealing with measurement of human traits and attitudes, etc....etc) indicate that any question can never achieve a max. error rate not lower than 80% (i.e. AN ERROR RATE OF 80% OR MORE IS ALWAYS UNACHIEVABLE), even after allowing for a max. grace variation of up to 5%. This means that no matter how difficult a question is, if it is presented to a well-represented respondent sample population of at least 1000, you will find that GENERALLY (IN MOST CASES) the toughest question will be answered wrong by at most 75% i.e. 750 people, not more. However, since the error rate is subject to fluctuations (as the number of respondents keeps increasing above 1000), it will still never exceed 79.99999......999%. (i.e. in any case, at least 201 respondents out of 1000 have to answer the question correctly, or else there is something wrong with the question). This proves that the answer to the new question has to be necessarily wrong as the error rate is given to be more than 80% and hence resolves (supports) the contention held by the respondents (referred to in the last line of the passage)
x-x-x-x-x-x-x-x-x-x-x-x-x-