CAT 2011 - Thoughts based on CAT 2010

CAT 2010 results were released a few days ago, and it is time to think about what we have learn from the results. I have categorized my thoughts into two categories.

Opinions that have been reinforced

1. CAT is not an impossibly tough exam. A number of diligent non-genius candidates have done really well. (All the evidence I am going to refer to is anecdotal and not based on any survey.). Thorough preparation, lots of practice and good planning should be enough to get candidates close enough. The final ingredient is perhaps a little bit of luck, but we cannot budget for that. The previous post on Self-belief holds good even now.

2. CAT rewards preparation from first principles: Quant, DI and verbal have all become more application-intensive and CAT 2010 has continued on with that trend. There is a higher bias towards non-formulaic questions. In maths and verbal, intuition and deeper-understanding is getting rewarded vis-a vis blind formulaic learning. As my boss never tires of saying - Intuition can be built with practice.

3. Balanced preparation is a must: With competition this high, one cannot afford to say my strength in quant should take me through. A few of our students learnt that lesson this time around

New pointers that CAT 2010 has shown us

1. Quant level across the country is pretty high: 15 years ago, \a student needed to just now a bunch of formulae, and need not have been conceptually sound. 6-8 years ago, when CAT made a shift towards more application-intensive questions, it was sufficient if one was conceptually sound. And you could get away without much practice. You always had time to derive 1-2 formulae, do trial-and-error and build hypotheses, verify with bunch of examples, etc. Now, the luxury to do all that is disappearing. A student almost needs to pick the right method to solve a question straightaway. No time for any trial-and-error business. One needs to have basics sound and practice gazillions of questions. The more different kind of problems you can lay your hands on the better.

2. To crack DI, one needs to be good at DI and LR: One out of two wont go. There are some tough DI passages and tough LR questions that you are better of leaving. The option of "I will kill DI and leave all LR questions" will not work

Those of you who are preparing for CAT 2011 and beyond, best wishes.

CAT - In defence of the CAT

As CAT 2010 draws to a close, we will soon be hearing a lot of feedback on the quality and consistency of the exam, the technology, organization, etc. It is about the same time that we will hear people saying that the CAT is a pointless exam, an exam that tests unnecessary skill sets, and an exam that is needlessly too difficult. People will tout examples of various calculation-intensive questions friends of friends of theirs saw in the paper and feel vindicated about their claim that the CAT is a very computation-heavy random exam designed to test unnecessary things. You will probably also hear the inevitable "Why is it necessary to do 24.5 * 35.4 in 10 seconds to be a good manager".

Most of the people who say this know little to nothing about the CAT, and at any rate have not spent enough time either thinking about the paper or evaluating it. They are happy to accept the established stereotype and help it along.

My colleagues and I have been taking the CAT almost every year and can state unequivocally that the quality of the paper has improved with time. There are 3 key myths that people have been spreading about the CAT

1. It is a calculation-intensive exam: On an average, about 5 out of 40 questions in Quant + DI are of this type, and one can skip this and still score 100th percentile in CAT. CAT was calculation-intensive in the late 90's, it has tested application heavily in the past 10 or so years.

2. CAT does not test relevant skills to be a manager - also stated as "Why should someone know Set theory to be a good manager?": At some level, CAT is an entrance exam designed to test basic intelligence. Any academic performance is a signal that the student can potentially do well. So, the CAT is looking for a signal of intelligence. Around two lakh students of diverse backgrounds take this exam every year. One needs to find something very basic to use it as a proxy for intelligence. Testing numerical ability, problem solving ability and comprehension ability are probably the best proxies available. What would you rather test - general knowledge, science, subject-knowledge? Anything else pales in comparison. Also, remember, CAT is not an entrance exam to be a manager. It is an entrance exam to get admission into a school that will train you to become a manager.

3. Some questions in CAT are too difficult: This is a problem that has come about because of a specific characteristic of India - one of large numbers. CAT needs to make a distinction between the average and the good. But if this were the only distinction that CAT needed to make, a consistent and simple paper would suffice. But the CAT also needs to make the distinction between good and really good and really good and exceptional. Remember, they need to devise a mechanism to distinguish the top 0.2 percent within the top 1 percent. This suggests the need for creating "men and boys" questions. Questions that demarcate the exceptional from the merely very good. Any paper where the really good students can attempt 55+ out of 60 students within 135 minutes clearly indicates that this is not too difficult. If the paper is peppered with simple questions, the exam just tests just speed, and not understanding. A slightly tougher paper that requires a high level of application is required for getting the top 0.2% from a sea of applicants.

Most people will agree that it does not require a genius to crack CAT. It just requires loads of application, good decision-making ability and adequate preparation. Sounds like a good test for finding managerial talent to me :)

Now, finally to the point that irks people the most, why should calculation-speed be a factor at all? When the whole globe does not really rate calculation-speed, why does India (CAT) cling to this notion? The answer again, is straightforward. India still believes it is a critical skill set. And the CAT is trying to find the best proxy to select smart Indian kids. Why should the CAT not test what India finds important. If anything, the CAT has been doing a lot to de-emphasise this feature. Over the past 10 years, the number of number-intensive questions have fallen sharply, and rightly so. India is moving on, and the CAT is setting the pace. CAT cannot move on all on its own, then they will end up getting all the wrong candidates.

So, there goes my defence of the CAT. I am sure the IIMs wont care much about giving their viewpoint. So, here is my version of how they might have done it.

Having said all this, some things irk me a lot. The complete lack of consistency in the paper is an issue that CAT needs to address. More importantly, the CAT needs to be error-free. Especially now that they have multiple sessions. The CAT was not error-free this year. Far from it. And god knows how they treat the wrong questions. I cannot think of a single fair way of treating error-filled papers. Hope they improve on this.