Last visit was: 19 Nov 2025, 08:58 It is currently 19 Nov 2025, 08:58
Home
GMAT
Messages
Tests
Schools
Events
Rewards
Chat
My Profile
Close
GMAT Club Daily Prep
Thank you for using the timer - this advanced tool can estimate your performance and suggest more practice questions. We have subscribed you to Daily Prep Questions via email.

Customized
for You

we will pick new questions that match your level based on your Timer History

Track
Your Progress

every week, we’ll send you an estimated GMAT score based on your performance

Practice
Pays

we will pick new questions that match your level based on your Timer History
Not interested in getting valuable practice questions and articles delivered to your email? No problem, unsubscribe here.
Go to My Error Log Learn more
Close
Request Expert Reply
Confirm Cancel
User avatar
Tarmac
User avatar
Current Student
Joined: 18 Oct 2007
Last visit: 12 Sep 2011
Posts: 431
Own Kudos:
71
Given Kudos: 2
Location: USA
Schools:Tepper '11
Posts: 431
Kudos: 71
Another Cheating Scandal 02 Jul 2008, 23:23
Kudos
Add Kudos
Bookmarks
Bookmark this Post
I don't see how cheating can be avoided on a test which is given in 500 locations around the world, 3x a day, every day of the year.

GMAC is greedy in encouraging people to take and retake at $250 a pop. If you don't have the means to write 10 million unique questions, you must reduce the frequency of tests given.

I believe the bar exam is given 2x a year, it is near impossible to cheat on it.
User avatar
filmcity
User avatar
Current Student
Joined: 01 Aug 2007
Last visit: 27 Oct 2010
Posts: 156
Own Kudos:
13
Given Kudos: 1
Posts: 156
Kudos: 13
Another Cheating Scandal 02 Jul 2008, 23:32
Kudos
Add Kudos
Bookmarks
Bookmark this Post
kryzak
kidderek, I feel your rage and understand it. Trust me, this is NOT the first time I've heard of the rampant cheating going on with the GRE's in China and India and why the GRE's are back to paper format in those countries again.
Show more

I wrote both GRE and GMAT and it is painful to read these blanket statements. In India GRE is very much online. I do not know about China, but regarding India the picture you people are painting is exaggerated. Yes, you'll show me the scoretop visitor trend but the thing is that, that was probably the only illegitimate avenues for Indians ( since i)they do not have the comfort of their own common vernacular websites and ii) even in english, I have not heard of ANY single suspicious test preparation website run by an Indian, I may be wrong here, please share ino if I am wrong ).Of course that does not take away the culpability of the VIPs, but THIS fact do help in moderating the grim scenario people are imagining . And although many Indians were visitors to the site but not soooo many were paid members ( but Here I agree that probably among paid members also Indians may outnumber other, but that is just the reflection of the applicant demographic). People saying that scoretop never came in google search, then I think they were using some different google.

I am sure if you search for similar trends of other GMAT websites, for most of the forums you will find majority visitors from India.

I prepared for GMAT almost in isolation in virtual world without much interaction with people writing it but I wrote GRE while I was still in college, where many people were writing the test, and I have not seen/heard anyone having access to any real question. There are many Indians in this site, I request them to chip in with facts or their first hand knowledge, if they think otherwise( I have heard many things like "my friends told me while I visited" or so, but none of my friends have ever told me and I have spent my entire life in India; I pity on myself, which kind of dull friends I have been cultivating for so long)

This country-wide name calling just kills me and I try to remain aloof, but sometime it just reaches nadir. Sorry, if I have spoken too much. :(
User avatar
bhatiagp
User avatar
Joined: 01 Jan 2008
Last visit: 06 Jan 2021
Posts: 371
Own Kudos:
116
Concentration: General Management , Strategy
Products:
My Reviews
Posts: 371
Kudos: 116
Another Cheating Scandal 03 Jul 2008, 00:55
Kudos
Add Kudos
Bookmarks
Bookmark this Post
The GRE is not paper based in india. Regarding cheating on the GRE in India, I doubt that happens . Out of 600 students in my Engineering class, more than 70% of the students gave their GRE. Its the same situation year after year.The only incorrect thing that probably happened was photocopying the ETS GRE Big Book because the original wasnt available then in India, and needed to be ordered through Amazon , and sharing the Kaplan, Princeton tests. Doubt anyone had purchased the books too for the cds..

Apart from that with so many students taking the exam year after year there was enough awareness on how to prepare , what to study etc, and from where to study. Dont think any of my classmates use the internet too for their GRE preparation. It was Barons, GRE Big Book, and the Kaplan, Powerprep , and Princeton tests. I can only speak for the people I came across, but its a sizeable number, and if there was an avenue where live questions were discussed, someone would have known, and the word spread. I remember that there were some Chinese sites from where one can download software based tests, from Kaplan, Princeton, ARCO etc.

I think Indians have been doing well on the GRE and GMAT primarily because of their strong quantitative skills. On the Verbal section they probably slog more than the others, owing to the fact that the English tested is American and we still use the British English. Those who hone their verbal skills invariably do well on the test.

Again its a well known fact that less than 1200 score /1400 or 2000/2400 on the GRE for an Indian is not considered good by the ADCOM's , and because not many can afford to pay for repeat tests, they generally work longer and harder towards these exams. Regarding the GRE many take up exams in the final year of their colleges when theyre still studying, with their peers, so it again helps in boosting scores, as they study in a group, and their quant skills are in practice. This reflects the majority of the students. Its not cheating that helps them achieve good scores but the overall effort they put in towards the exam in general, and the exam knowhow , specially in a college where majority of the students take the examination.
I remember so many students postponing their exam because they were unable to get that 2000+ score in their practice.

Again with a large student population, it is obvious that more number of people as compared to other nationalities get higher scores. If you look at it, theres a popular myth in India that its the score that gets you the admit. This kills most applicants, as they tend to have great scores, but poor EC and written applications. This has been the case since ages.

Let us not mix the GRE and the GMAT, as the represent different set of applicants. The scoretop issue is extremely different, and does not reflect the candidates who sit for the GRE. Incase someone differs, please update me.. I may not be up to date with current trends then....
User avatar
kryzak
User avatar
Major Poster
Joined: 04 Jun 2007
Last visit: 10 Aug 2013
Posts: 5,452
Own Kudos:
744
Given Kudos: 14
Status:Um... what do you want to know?
Location: SF, CA, USA
Concentration: Technology, Entrepreneurship, Digital Media & Entertainment
Schools:UC Berkeley Haas School of Business MBA 2010
GPA: 3.9 - undergrad 3.6 - grad-EE
WE 1: Social Gaming
Posts: 5,452
Kudos: 744
Another Cheating Scandal 03 Jul 2008, 01:20
Kudos
Add Kudos
Bookmarks
Bookmark this Post
All I'm relaying is what I've heard from friends in Engineering grad school (Indian and Chinese international students) regarding the GRE (not GMAT and not scoretop related). Obviously I don't know it first hand and I am not saying it as a blanket statement. I apologize if it read that way and offended anyone. I'm using the similar issues I've heard about the GRE and the action they took (probably only in China, since people say India it's still CAT) as a comparison to what GMAC should do about the GMAT.
User avatar
bhatiagp
User avatar
Joined: 01 Jan 2008
Last visit: 06 Jan 2021
Posts: 371
Own Kudos:
116
 [1]
Concentration: General Management , Strategy
Products:
My Reviews
Posts: 371
Kudos: 116
 [1]
Another Cheating Scandal Updated on: 03 Jul 2008, 03:37
Originally posted by bhatiagp on 03 Jul 2008, 01:38.
Last edited by bhatiagp on 03 Jul 2008, 03:37, edited 3 times in total.
Kudos
Add Kudos
1
Bookmarks
Bookmark this Post
kryzak : Its fine.. I didnt feel offended but I would ifeel bad if Indians cheat their way up and take deserving places from other candidates, who have worked hard , as such behaviour when exposed reflects negatively upon the entire commity..

Heres some food for thought. This has been troubling me since I read about how the live questions are collected.

Look at it, does it make business sense to pay a candiate ( Few hundred $) to write the test, pay 250$ for his registration and just get 1 or 2 questions instead from him, again your not sure if the question is remembered exactly by the candidate. Remember the site sells 400 questions a month, which means its spending atleast 120,000$ per month . (300$ per candidate *400 (One question per person) ) This means atleast 4000 active VIP members every month are needed to break even, at 30$ each. Im sure paypal would also be charging per transaction, which means its less than 30$ that they effectively get from a member. Adding the cost of employees etc, who would compile these documents, instructors, who would solve these problems and post the answer, etc to the above cost will make their overall cost very high.

Secondly I doubt that makes logical sense to pay someone for remembering just 1 or 2 questions , especially because the GMAT tests different areas, and it is not easy to remember CR and RC's, or long word problems. Again remembering punctuation marks like semi colons, and commas etc in sentences is not as easy, as we think it is. For e.g , read a SC question on Verbal ,and try writing it down after 1.5 hour , and see if you remembers all 5 options correctly. Its not possible

Limiting to questions that their paid test takers collect will not help majority of the VIP members as they pay for practise across a variety of questions. It is also not possible to synchronize various test takers that are sponsored by such sites to collect only new questions and avoid collecting previously collected questions.

Im sure theres more to this issue than what meets the eye. I feel they maybe using a different strategy.
I speculate the following scenarios

1) Students may Leak Questions by writing them on scratch sheets
Firstly, I question if all the test centers are following common procedures. For e.q Many test centers dont maitain similiar standards across all countries. In many countries, writing pads are given for scratch work for the GMAT unlike in the US, where you are given an erasable board with marker. I have a strong feeling that there is a possibility, that someone can come for the test, write the questions on his scratch pad and carry some of the sheets with him.

2) Scratch Sheets not disposed or sold
I beleive it could not only be students leaking questions but also scratch sheets for tests are not effectively disposed or shred. In many poor countries , a few 30-40$ can equate to more than a test center employees monthly salary, and what does it take to sell these rough scratch books rather than disposing it.

3) ( Similiarity with brain dumps for IT certification)

There is an analogy here with IT Certification exams. I dont figure out on how sites create brain dumps (live question) pool for IT certification exams. Many IT companies dont stop this as running such exams are a revnue stream and it is better if more people site for these exams, pass these examsn and retake them again in two years to renew their certification .

Ive interviewed people who have no clue of the technology their certified in; but they were certified engineers by Microsoft, CISCO etc. All they did was study the brain dumps, and sit for the exam. Some cisco certified proeffsionals had never ever seen a router in practice. Imagine they were certified engineers.Request you guys to type brain dumps and see the number of sites that provide such material for exams ... Its even tougher to remember IT certification test questions as many questions in IT certification esams are extremely long, and some are pictorial which cannot effectively be redrawn unless someone has access to these questions. Unless someone can cut and paste this pictures onto word and then compile a document it is not possible. Maybe the same is happening with GMAC . I mean it only take an alt -tab to move between different windows and do a cut and paste operation. There is a possibility that the test center employees assist such some people by allowing these test takers to compile a list of questions and then emailing them.

I doubt GMAC has the time to watch every recorded session held across so many countries, to verify if cheating is taking place.

Knowing exactly how brain dumps are collected can probably help GMAT kill this problem.

But this brings me to another issue. Braindumps have been existing since ever since IT certification started which is much before GMAC moved its test centers from ETS centers to Pearson run centers in 2005. Didnt the know about the exisitng issues with IT certification exams, and the open availability of brain dumps for IT certification. Didnt they ever think if there is a role of certain test centers in leaking these questions.

My remedy

1. GMAT must move back to the ETS testing centers . This will probably solve most of the issue and atleast people wont have uncommon test experiences.

2. Administer the exams not more than four times a month with a different question pool for each exam.

3. Replace paper and pencil with erasable boards in centers not currently offering the erasable boards,

4. Give a limited time frame to answer every question, where the question changes automatically if not answered in stipulated time.

For e.g. GMAC can time every question based on the questions difficulty , for eg those in sub 600 range can be given 3 minutes maximum, 600-700 maybe 4 minutes and tougher questions like RC and tough CRs maximum of 5-8 minutes , with the test automatically moving to next question, this may solve this problem, as people may not have enough time to memorize a question and remember the answer choices.

This indirectly will again be a benefit to so many test takers, as they can decide on how much the time question deserves and it will improve time management amongst the actual test takers.
User avatar
riverripper
User avatar
Major Poster
Joined: 10 Apr 2007
Last visit: 20 Aug 2022
Posts: 4,306
Own Kudos:
806
Given Kudos: 5
Location: Back in Chicago, IL
Concentration: General/Operations Management
Schools:Kellogg Alum: Class of 2010
GMAT 1: 740 Q49 V42
Posts: 4,306
Kudos: 806
Another Cheating Scandal 03 Jul 2008, 03:29
Kudos
Add Kudos
Bookmarks
Bookmark this Post
Cheating happens in all places, more so in other countries than the US. I would say China is far and away the biggest location of GMAT cheating, India may be second because of sheer numbers of test takers and the desire to do extremely well. However, in China it is a HUGE problem and something needs to be done along the lines of the GRE. It definitely seems to be a cultural thing, it really doesnt seem like they think their is anything wrong with what they are doing...its not an excuse but its harder to blame people who dont really believe that its as wrong as we do.
User avatar
rhyme
User avatar
Major Poster
Joined: 05 Apr 2006
Last visit: 02 Dec 2024
Posts: 5,909
Own Kudos:
3,172
Given Kudos: 7
Affiliations: HHonors Diamond, BGS Honor Society
Schools: Chicago (Booth) - Class of 2009
GMAT 1: 730 Q45 V45
WE:Business Development (Consumer Packaged Goods)
Schools: Chicago (Booth) - Class of 2009
GMAT 1: 730 Q45 V45
Posts: 5,909
Kudos: 3,172
Another Cheating Scandal 03 Jul 2008, 05:06
Kudos
Add Kudos
Bookmarks
Bookmark this Post
Travel09

I think this is were problem is. We are reaching conclusions based on assumptions and imagination
Show more

I have to disagree....... He's being very fair. Based on all the evidence, I think its quite reasonable to conclude that the term JJ was used a lot on the site, and his "I can imagine" is simply a way of illustrating his point.

People said things like (and I quote): "I have posted around 30 MJJs, which I could remember, from my exam today in the 6.06 section." Moreover, in subsequent posts where someone doesn't report his or her test questions they are asked: "pls recollect and post the test questions", and someone else asks "Can you please post the JJs?" followed by "Did you see any Sep JJs?" (to which he later says he did - to the tune of 10 or so in math). In another post someone asks "I read several GMAT experiences and many ex-GMAT takers spoke about MJJs and VJJs.......but is it the real question"? (The reply is that it is). Moreover, The VIP forums were ordered by month, so you could get access to the "June" questions by going to the 06.2007 VIP subforum. The July questions were in the 07.2007 subforum. In other words, even the structure of the forum as a whole seems to have been put together for the purposes of organizing live questions into the appropriate month so users could find ones relevant to their planned test date (i.e. if you are taking the test in 07.2007, then one would look at 06.2007 but not 06.2006 or something)....

So when terp says, "I can imagine a thread titled something like " June JJ - Probability " " - thats not based on assumptions, nor is it concocted from thin air. It's based on a mountain of evidence that, to any reasonable individual, would seem to strongly support the possibility.

If a house is on fire, I don't need to go inside to know there's smoke.
User avatar
kidderek
User avatar
Joined: 24 Aug 2006
Last visit: 13 Jul 2019
Posts: 1,960
Own Kudos:
364
Posts: 1,960
Kudos: 364
Another Cheating Scandal 03 Jul 2008, 06:35
Kudos
Add Kudos
Bookmarks
Bookmark this Post
rhyme

If a house is on fire, I don't need to go inside to know there's smoke.
Show more

In this day and age, some people aren't convinced unless they see a grainy video clip of the incident. This is the same country that let OJ walk.
User avatar
terp26
User avatar
Current Student
Joined: 22 Oct 2006
Last visit: 06 Apr 2020
Posts: 1,210
Own Kudos:
380
Given Kudos: 12
Schools:Chicago Booth '11
Posts: 1,210
Kudos: 380
Another Cheating Scandal 03 Jul 2008, 06:36
Kudos
Add Kudos
Bookmarks
Bookmark this Post
For someone with a good memory it is extremely possible to remember 20+ questions.

They really need to come down on people for this. I can't help to think the value of my GMAT just diminished a little bit and it gets me a little angry. I do feel sorry for the Indian and Chinese members as their super high scores will probably be even more scrutinized and less-valued by AdComs. They will be very wary of people with high verbal scores especially. It is ashame.

As for them changing the format of the test, I highly doubt they would go that route unless they were threatened with their contract.

As a data person, I would imagine they have done numerous data mining efforts to discern abnormalities in scoring.

Example would be, they have a hit-rate/difficulty on a question in the beginning of the month, so they expect 25% of people to answer that question correctly, if that questions hit-rate deviates statiscally significantly upward towards the end of the month (When people would have access to the JJs for that month) They could easily capture the questions posted and any abnormalities. Now this is assuming the cheating is enough to have that much of an impact. However if they break it down by testing location, or notice scores in China are significantly higher than verbal scores in the US towards the end of the month, then they may be able to gauge the extent or type of questions that were cheated on.

Obviously with such a large testing pool, it may just be a blip on the radar , but they may be able to determine such things statistically (but hey, that is a reason why I want to go to b-school to learn)
User avatar
kidderek
User avatar
Joined: 24 Aug 2006
Last visit: 13 Jul 2019
Posts: 1,960
Own Kudos:
364
Posts: 1,960
Kudos: 364
Another Cheating Scandal 03 Jul 2008, 06:47
Kudos
Add Kudos
Bookmarks
Bookmark this Post
Not sure if it's been mentioned, but practicing with "live" questions gives someone an edge over those who practice from just OG/GMATPrep. These "live" questions, though they may not appear on the test, are the closest thing to the questions that would appear on the test.

Considering we all know the importance of OG & GMATPrep, it's a significant edge.
User avatar
Audio
User avatar
Current Student
Joined: 13 Jun 2007
Last visit: 31 Jul 2012
Posts: 408
Own Kudos:
87
Concentration: Finance
Schools:Wharton, Booth, Stern
Posts: 408
Kudos: 87
Another Cheating Scandal 03 Jul 2008, 09:50
Kudos
Add Kudos
Bookmarks
Bookmark this Post
Tarmac
I don't see how cheating can be avoided on a test which is given in 500 locations around the world, 3x a day, every day of the year.

GMAC is greedy in encouraging people to take and retake at $250 a pop. If you don't have the means to write 10 million unique questions, you must reduce the frequency of tests given.

I believe the bar exam is given 2x a year, it is near impossible to cheat on it.
Show more

There is an efficient and easy way to force GMAC to do something about it: MBA schools should simply boycott the test and create / develop one of its own, or adopt another test by another company.

Concerning cheating, I have mitigated views on the whole scoretop issue. I'm no law expert, but I really don't see how Gmac is going to prove which VIP members really took advantage of the system and which not (yes I do think that some of those VIP members are innocent, despite the fact that ignorance is no excuse). I really hope that I'm wrong, but it's going to be difficult I reckon.

I agree with another poster concerning punishments (sorry, forgot the name): all VIP members should have their scores cancelled (not a fair view of their abilities), and people who knew what they were doing should not be allowed to represent their exam.

I do hope that the situation is somehow going to change though, because this is perfectly unacceptable for all the honest test-takers. I'm sure that some of these honest people have lost their place in the B Schools they were applying for to some of these cheaters!
User avatar
Audio
User avatar
Current Student
Joined: 13 Jun 2007
Last visit: 31 Jul 2012
Posts: 408
Own Kudos:
87
Concentration: Finance
Schools:Wharton, Booth, Stern
Posts: 408
Kudos: 87
Another Cheating Scandal 03 Jul 2008, 10:03
Kudos
Add Kudos
Bookmarks
Bookmark this Post
riverripper
Cheating happens in all places, more so in other countries than the US. I would say China is far and away the biggest location of GMAT cheating, India may be second because of sheer numbers of test takers and the desire to do extremely well. However, in China it is a HUGE problem and something needs to be done along the lines of the GRE. It definitely seems to be a cultural thing, it really doesnt seem like they think their is anything wrong with what they are doing...its not an excuse but its harder to blame people who dont really believe that its as wrong as we do.
Show more

I can't believe people don't think that they are doing is not wrong. They either don't understand the situation, don't know what the test is or are lying to themselves. Therefore I don't believe that it's a cultural thing. I just think that it's due to the fact that these people are desperate to do well and get themselves out of poverty, and most importantly, they have the logistical possibility to cheat thanks to their sheer number.
User avatar
kidderek
User avatar
Joined: 24 Aug 2006
Last visit: 13 Jul 2019
Posts: 1,960
Own Kudos:
364
Posts: 1,960
Kudos: 364
Another Cheating Scandal 03 Jul 2008, 10:05
Kudos
Add Kudos
Bookmarks
Bookmark this Post
Audio

There is an efficient and easy way to force GMAC to do something about it: MBA schools should simply boycott the test and create / develop one of its own, or adopt another test by another company.
Show more

So true, yet so not going to happen.

Audio

(yes I do think that some of those VIP members are innocent, despite the fact that ignorance is no excuse).

I agree with another poster concerning punishments (sorry, forgot the name): all VIP members should have their scores cancelled (not a fair view of their abilities), and people who knew what they were doing should not be allowed to represent their exam.
Show more

Actually, ignorance of fact (thinking that these are legit questions) is an excuse. However, the scores are tainted nonetheless. The test takers who saw "live" questions innocently should be allowed to retake; those who knew that they had "live" questions should be banned and expelled.

Audio

I do hope that the situation is somehow going to change though, because this is perfectly unacceptable for all the honest test-takers. I'm sure that some of these honest people have lost their place in the B Schools they were applying for to some of these cheaters!
Show more

It sucks. Everyone gets a black eye for this fiasco.
User avatar
IanStewart
User avatar
GMAT Tutor
Joined: 24 Jun 2008
Last visit: 18 Nov 2025
Posts: 4,145
Own Kudos:
10,988
 [1]
Given Kudos: 99
Expert
Expert reply
Posts: 4,145
Kudos: 10,988
 [1]
Another Cheating Scandal 03 Jul 2008, 10:13
1
Kudos
Add Kudos
Bookmarks
Bookmark this Post
terp26

As a data person, I would imagine they have done numerous data mining efforts to discern abnormalities in scoring.

Example would be, they have a hit-rate/difficulty on a question in the beginning of the month, so they expect 25% of people to answer that question correctly, if that questions hit-rate deviates statiscally significantly upward towards the end of the month (When people would have access to the JJs for that month) They could easily capture the questions posted and any abnormalities.
Show more

I thought this would make an interesting though experiment- to see if GMAC could have worked out, from data mining, whether any questions were abnormal. Based on a few assumptions, I arrived at the following (rather long, and possibly completely uninteresting to most!):

  • Roughly 20,000 people take the GMAT each month. Half of these, 10,000, take the test in the second half of the month. We'll only look at them.
  • 6000 people were allegedly ScoreTop subscribers. Likely more benefitted from the availability of 'live' questions, but I'll go with the assumption that 6000 benefitted for now.
  • I think ScoreTop was distributing live questions for four years, though am not sure. In any case, using that assumption, 1500 test-takers would have seen, in advance, live questions each year, or 125 per month. I'll assume all of these took their test in the second half of the month.
  • According to some of the propaganda used by ScoreTop to advertise the value of their question sets, there are roughly 1200 questions in the GMAT test pool in a given month. It's unclear whether that's per section, or in total, or even whether that number is at all accurate (GMAC does not disclose the size of their question pool). I'll use that as an assumption- that there are 600 questions in the question pool for each section of the test.
  • Then each test-taker sees about 6% of the total question pool on each section. I'll assume there is an equal probability of seeing any given question- not a valid assumption, but I need to make the analysis simple.
  • For someone to see about 7-8 questions from JJs repeated on their real test (as some bloggers report), ScoreTop must have had roughly 20% of the question pool available in the JJs.
  • Now, to analyze how you would expect test-takers to perform on a given GMAT question, it's important to understand how the CAT works. The test is designed to adjust the difficulty of questions until it reaches a point where test-takers answer 60% of questions correctly, 40% incorrectly (I'd note that some sources cite the figure at 50%, which is only correct for a test without answer choices; the GMAT accounts for the possibility of guessing correctly). I'll take this as a baseline- a legit test-taker should answer a question correctly, mid-test, with about a 60% probability.
  • Suppose JJ-users answer previously seen questions with 100% accuracy.
  • By the assumptions that each test-taker sees 6% of the pool, that we have 10,000 - 125 = 9875 legitimate test-takers, and 125 JJ users, and that each question is seen by an equal number of test-takers, we arrive at the following: a given QN that appears in the JJs will be shown to approximately 0.06*9875 = 592.5 legitimate test-takers, and 0.06*125 = 7.5 JJ users, and to 600 people in total.
  • One would expect 360 correct answers to such a question. Instead, one would receive 0.6*592.5 + 7.5 = 363 correct answers.
  • By binomial probability, the standard deviation around 360 is sqroot(600*.6*.4) = 12. That is, if there were no cheaters, one would expect, 68% of the time, that the number of correct responses would vary from 348 to 372 on a given question. For the number of correct answers to be different by only a quarter of a standard deviation is immaterial as statistical evidence- you'd see, on a completely fair question, 363 or more correct answers 40% of the time.

So my guess is that it would be difficult to identify problematic questions by data-mining. I've needed to make a lot of assumptions, however.
User avatar
terp26
User avatar
Current Student
Joined: 22 Oct 2006
Last visit: 06 Apr 2020
Posts: 1,210
Own Kudos:
380
Given Kudos: 12
Schools:Chicago Booth '11
Posts: 1,210
Kudos: 380
Another Cheating Scandal 03 Jul 2008, 10:18
Kudos
Add Kudos
Bookmarks
Bookmark this Post
nice analysis ! need to re-read to digest everything +1
User avatar
bhatiagp
User avatar
Joined: 01 Jan 2008
Last visit: 06 Jan 2021
Posts: 371
Own Kudos:
116
Concentration: General Management , Strategy
Products:
My Reviews
Posts: 371
Kudos: 116
Another Cheating Scandal 03 Jul 2008, 10:19
Kudos
Add Kudos
Bookmarks
Bookmark this Post
Hey Ian

Your quant skills are impressive.. Hope to see you more on the Math forum..
User avatar
sam77sam7
User avatar
Current Student
Joined: 02 Jan 2008
Last visit: 06 Mar 2012
Posts: 597
Own Kudos:
30
Location: Detroit, MI
Concentration: Consulting
Posts: 597
Kudos: 30
Another Cheating Scandal 03 Jul 2008, 10:58
Kudos
Add Kudos
Bookmarks
Bookmark this Post
Ian,

Good analysis. Once you really break down the numbers, for 6,000 out of the 250,000 a year that take the test, their influence was probably pretty minimal...

Good work, kudos!

~Sam
User avatar
kidderek
User avatar
Joined: 24 Aug 2006
Last visit: 13 Jul 2019
Posts: 1,960
Own Kudos:
364
Posts: 1,960
Kudos: 364
Another Cheating Scandal 03 Jul 2008, 11:05
Kudos
Add Kudos
Bookmarks
Bookmark this Post
IanStewart,

Great attempt. As you've said, I think there are way too many assumptions. But, then again, you don't have much else to work with. Considering only 8% or so score above 700, the number of 700 level questions greatly reduces the field. I also think that there are far fewer questions in each difficulty bin than most people think.
User avatar
jasonc
User avatar
Current Student
Joined: 18 Jun 2008
Last visit: 09 Sep 2011
Posts: 199
Own Kudos:
79
Given Kudos: 3
Posts: 199
Kudos: 79
Another Cheating Scandal 03 Jul 2008, 11:09
Kudos
Add Kudos
Bookmarks
Bookmark this Post
sam77sam7
Ian,

Good analysis. Once you really break down the numbers, for 6,000 out of the 250,000 a year that take the test, their influence was probably pretty minimal...

Good work, kudos!

~Sam
Show more

Sam, I think you may be reaching the wrong conclusion.
Ian was breaking down the numbers under a number of assumptions (as he stated) to understand if its possible to identify problematic questions by data-mining, NOT to approximate the potential impact of this scandal.
Ian understood the flaws in his assumptions well but I wanted to emphasize so people don't underestimate the potential impact of cheaters
Quote:
Then each test-taker sees about 6% of the total question pool on each section. I'll assume there is an equal probability of seeing any given question- not a valid assumption, but I need to make the analysis simple.
Show more
User avatar
sam77sam7
User avatar
Current Student
Joined: 02 Jan 2008
Last visit: 06 Mar 2012
Posts: 597
Own Kudos:
30
Location: Detroit, MI
Concentration: Consulting
Posts: 597
Kudos: 30
Another Cheating Scandal 03 Jul 2008, 11:15
Kudos
Add Kudos
Bookmarks
Bookmark this Post
Jasonc,

Sorry about that, I reread the post. You are correct.

However, I would still make the case that the influence of the 6000 people's scores on the remaining test takers score is probably pretty minimal. I don't know how GMAC calculates it all though, nor do I have the time (or possibly the ability) to do the same type of analysis with regards to cheaters impactl.

~Sam
   1  ...  15   16   17   18   19  ...  22