Search This Blog

Sunday, January 18, 2015

History of IIT JEE

Someone asked me to answer a question on how JEE has changed over the last few years. That inspired me to write about the last 20 odd years that I have been at IIT Kanpur, and have seen the JEE process from close quarters. Of course, I could write about JEE of earlier times as well, but this was becoming too long and I don't have as authentic information for the earlier years, and hence I am restricting myself to the last 20 years.

Till 1996, we had three exams over two days, Physics, Chemistry and Maths. Each exam were of three-hour duration. All three would be graded, and the total would be used for ranking. In 1997, we had the infamous JEE paper leak. We had to conduct the exam again a month later. A lot of meetings, discussions, and arguments later, it was agreed that securing an exam for two lakh students (and growing) over 2 days is very difficult. Even before this paper leak, there were rumors that some ghost writers were giving JEE for others. Also, there were murmurs that grading almost 2 lakh copies in a consistent way was getting difficult despite a very elaborate process that IITs had to ensure consistency of grading. And therefore, we must go for a two stage process where both stages were one-day affairs, first stage was objective type, and only a few people would give the second exam, which would be long answers. It was easy to secure an exam where there were only 20,000 candidates, and also ensure consistency in its grading. The screening test was conducted in the first week of December. The result was announced around 1st Jan. The final exam was held in first week of May (which was the traditional date for JEE till then), and the final result was announced around 1st June. This was supposed to de-stress students. Those who would know that they are not getting into IITs on Jan 1st, would still have 8-10 weeks to prepare well for their board exams. Most good engineering colleges (including NITs, BITS, DCE, etc.) admitted students on the basis of 12th class marks, and what was happening earlier was that students studied both for JEE and 12th class at the same time, and did not do well in either. This new arrangement was expected to allow students to focus on 12th from January onwards, and hence considered very student friendly. Also, we believed that if there is a significant gap between the results of prelims and the final exam, then the coaching will primarily focus on prelims (much larger market) and final exam would see less negative impacts of coaching. On the other hand, if prelims and final exams were close to each other, then the coaching would focus on the final exam and hope that this also helps in the prelims.

This new pattern was introduced in 1998. At some point in time, the Ministry of HRD got into the act, and started claiming that they have received complaints from several principles of schools that once the students know the result of JEE prelims and if they are not shortlisted, they get depressed and spoil their 12th class exams even more than what they would have done if they were studying for 12th class and JEE simultaneously. We never believed that story. We were quite sure that if we were only admitting 4000, and we were shortlisting 20,000 then the guys who were not short-listed would be mostly those who would actually be expecting not to be short-listed and could not be depressed by knowing the result. They should be relieved and not depressed. (A few may be depressed, but the policies can not be made for few. One has to look for the larger good.) And more importantly, this was our feedback from school teachers and principals though we must admit that MHRD has access to a much larger number of teachers and principals than us. But the argument never made sense to us. We always believed that it was personal agenda of someone in the Ministry, perhaps with some encouragement from coaching industry, but one will never know.

But another development taking place around that time was that many private universities (deemed universities) started having their own admission tests (initial focus was perhaps profits in conducting that exam, since government was becoming tough with having capitation fee, and large application fee, admission fee, etc. And Supreme Court was telling all universities and colleges that merit must be respected in the admission process. So having entrance exams solved both issues - making money and be on the right side of law. Many state governments also created technical universities and affiliated all engineering colleges of the state to this single university, and the admission will happen through yet another common entrance exam. So one common exam in each state started happening. As admissions became more competitive, the issue of normalizing board marks also started raising its head. This problem was too solved by having an admission test. But having all these admission tests meant that our original argument of 1997 that a student who is told on 1st Jan that he can not get into IIT will now focus on school board exams was becoming weaker every year. In reality, the child will just start preparing for the other competitive exams.

So finally, we shifted the prelim from December to 2nd week of April, would declare the result by 1st May, and then the final exam will be 10-15 days later. I don't remember which year it was that we shifted the prelims from December to April.

In 2001, the central government decided that there should be only one common entrance exam for all engineering institutes in the country. The idea of AIEEE was born. However, every body rejected the idea. IITs rejected the idea. They would continue to hold their JEE. Most of the deemed universities who were having their own entrance exams rejected the idea. Most of the state governments rejected the idea and continued to have their own common tests.

The government changed in 2004, and by now, the brand value of IIT system had become really big. So a new government and the ministry had to show that they knew how to run IITs better than the previous minister and the babus. There were genuine concerns that there were too many exams happening in the life of a 12th class student. Can we reduce the exams by one. From 2 JEE exams, can we make it one. Never mind that the second exam was given only by 20,000 students and removing that would have no impact on the number of exams by students at large. The ministry has to throw its weight and fool public. And the directors are all too eager to oblige in such situations. So we set up a committee. The committee looked at the data from JEE prelim and JEE final over the last few years and reported that the top 5000 in the two exams had more than 4000 common names. It said that the ordering of those 4000 was different in the two exams but that was considered statistically insignificant. Even if the same group give same exam on two different days, the ranking would be different, they argued. And, if there is very high correlation between the prelims and the final exam, why should we have the prelim exam at all.

There were many flaws in the argument. In fact, it was clear that everything mentioned was quite stupid. You don't find correlation by looking at how many people are common in the top 5000 and then ignoring the fact that their rank correlation was rather poor. Also, as it turned out, by removing the long answer type part, the whole coaching technique changed to solving the paper by elimination and tricks, and a lot more students started coming to IIT campuses who could not write a single sentence correctly (in any language, if I may add). So, having an exam that required long answers was ensuring that students learned more of desirable skills which was important even if there was strong correlation between the results of the two exams. (And in this case, the correlation was weak.) We also forgot that the idea of 2-stage process was to ensure security of the exam. We could check each candidate much more carefully when the exam was being conducted only for 20,000 students. We could require them to have stronger authentication. We could hold their exams in much better environment. Only good centers would be selected. You wouldn't lose out because the school did not have a diesel generator set and sufficient diesel to run that fan in your room while giving the exam. You wouldn't lose out because the desk you are sitting on is creaking and is making a noise. Later, we will also start receiving many complaints that there are ghost JEE takers whom we were unable to catch.

But the goal of JEE has not been to admit good students. That has been the side effect. The goal is to have a system which can stand the scrutiny of law and is apparently fair (any unfairness which is incidental and not planned is acceptable). If any system has these qualities and can also please the masters, that would be great. We got all these things right when we announced that from 2006, we would only have objective type questions in JEE, and there will be no preliminary stage. The masters were also getting worried that with almost all engineering colleges shifting to admissions through an entrance exam (AIEEE or something else), the focus on school education was getting lowered, and they wanted IITs to help. And the Directors obliged. The new rule about 60% requirement in 12th class was also added in 2006. (You can see that sometimes the Directors act very smartly. 60% was so low that in any given year, perhaps 1 or 2 in the entire IIT system would be disqualified as a result. So they could, in this one case, managed to please the masters with absolutely no impact on anyone.)

This made no dent on the problem, which was as expected by some of us. The students continued to give a large number of exams. They continued to ignore school education. In the meanwhile, another general election, another Minister, another set of babus. They too thought that they knew how to run IITs better, and the easiest thing to tinker with is the admission process, following the path of their illustrious predecessors. Another attempt at reducing the number of entrance exams, not realizing that "One Nation, One Test" slogan is just a slogan. The central government had no real authority to stop deemed universities or any state or private university from conducting their own exam. So they did two things, tried to merge AIEEE and IIT JEE, at least one major exam less, and tried to harass deemed universities through a variety of means so that they become more compliant and agree to stop their own exams (amongst other things). But it did not work. The number of exams that a student gives today is largely the same. But the pressure caused some restructuring. In order to show that the government has succeeded in reducing the exams, they decided that both the exams will be named JEE and one hoped that public is foolish enough to believe that JEE (Mains) and JEE (Advanced) are just two sessions of the same one exam, and somehow require significantly less preparation and effort than AIEEE and JEE. But the restructuring and renaming did mean that JEE (Advanced) which was the erstwhile IIT JEE would be taken by only top 1.5 lakhs of JEE (Mains), the erstwhile AIEEE.

Even though the number of 1.5 lakhs was decided as 5 times the number of CFTI seats (including all IITs, NITs, and IIITs), when NITs and IIITs decided not to take students from JEE (Advanced), and the number of seats reduced to 10,000 (from 30,000), no one in the IIT system had the guts to relook at this number and say that the advanced exam should be of 50,000 students alone (5 times the number of seats), which would allow us to have long answers. Having had JEE machine graded for almost 10 years now, it would be very difficult to find examiners for long answers. We have realized for the last 10 years that multiple choice questions reduce the number of court cases, and that is certainly more important than figuring out who is the best student to admit. The focus of IITs is also shifting to research. So if some good students go to NITs, it will be good for IITs. The chances of them coming to us for doing research is higher than our own students staying back to do research.

What is the future of JEE? Well, I am sure one day some committee will study JEE (Mains) and JEE (Advanced) and declare that both exams would have selected the same set of students for an IIT, and hence JEE (Advanced) should be abolished. There will be huge tussle between CBSE and IITs as the money involved is huge and the potential risk of damage of reputation is high if something does not happen right with the exam.


ksdjfksdf said...

I think 1998 was still in the old format - over 2 days.

It might have changed in 1999.

Manish Kumar said...

1998/1999 had two days(three hour papers) like 1997 and before. In 2000 Screening(3 hours) was introduced, It happened on 2nd Jan 2000 and result coming out on 22/23 Jan 2000!!
Those who got selected in screening again wrote 2-hours three paper in single day on 7th may. One other rumor(or truth?) very popular was whenever IIT have slight doubt of cheating/leak/ghost-writer, they cancel the whole center exam without public notice.

iitmsriram said...

"There will be huge tussle between CBSE and IITs as the money involved is huge". Money involved is not that big for the current JEE (advanced), net is only some 3 crores per IIT (in 2014, we had less than 1 lakh fee paying candidates and around 20% of these would be half fee paying, so gross revenue would have been about 18 crores). And, I don't know if IITs will want to the run the exam for 10 lakh+ candidates of the JEE (main), which is where the big money is. Last year we had about 13 lakhs candidates in JEE (main), gross revenue would be around 200 crores. This is bigger than GATE which nets maybe 20 crores per IIT.

gautam barua said...

I am hearing that there will be common counselling for jee(main) and jee (advanced) this year. Also that the 80 percentile cut-off for jee(advanced) will be replaced by 70% (?). Have you heard anything similar? What is your take on the implication of common counselling, if it happens? Besides the immediate impact, do you think it will be the beginning of the end of JEE(advanced)?

Dheeraj Sanghi said...

@Sriram, I agree that the amounts are not very large for an IIT.

Dheeraj Sanghi said...

@GB, I support joint counseling, since I consider this as a better thing than the current process. But to expect this to solve many problems, is expecting a bit too much. If you see how many of the 10,000 students selected for IITs also keep an NIT seat till the last minute and then select an NIT. (Because if s/he had selected an IIT, the seat would have been filled by NIT through future rounds, including a spot round. Also, the maximum seats vacant after the three rounds of counseling are in ISM Dhanbad, which I believe also does a spot round.) So the number of seats that remain vacant due to this problem are at best a couple of hundred, which is any way handled by institutes declaring a slightly larger number of seats than what they would like. In fact, the problem can be solved more easily by increasing the cost of keeping two seats from Rs. 1,000 to something like Rs. 10,000. No one, not even the rich, would unnecessarily keep extra seats if they stand to lose 10K for doing the same.

And we do have potential problems. What if the board results are not announced in time. NITs always succumb to pressure of filling seats after classes start. If there is joint counseling, would there be pressue on IITs too to do the same. So we need to brace for more political pressure on the IIT system.

There is a small advantage to students that they will have to login to only one portal, again not a big deal.

I see the biggest advantage as establishing the best programs of NIT system as equally popular (if not more) compared to the worst programs of the IIT system. Having such data would give confidence to more and more students to choose NITs over IITs, which will help reduce the gap between the perception of the two places, which can only be good for the society.

Dheeraj Sanghi said...

@GB, For 2015 admission, a student has to get either 80 percentile or 75% marks (which ever is lower) to be eligible for admission to an IIT.

I do think that eventually there will be political pressure to be similar to NITs in the admission process. Once the data conclusively shows that the architecture program in IIT Kharagpur is liked by less persons than Computer Science program of NIT Trichy, it would be difficult to argue that we are somehow much superior to NITs. So if NIT Trichy can admit students to its CSE program through JEE (Mains) and 12th class marks, why can't IITs do the same.

And I don't think the IIT faculty is ready for another fight so soon on the admission issue.

gautam barua said...

In 2014, Board results got delayed and the first round of seat allotments for JEE(Mains) took place on July 14, after the second round (or also third?) of allotments for IIT-JEE was over. Actually this worked well as students were no longer holding on to both an IIT and a NIT seat (IIT allocation having completed by July 14 or soon after). But then the counselling Board of JEE-Main allowed candidates not satisfied with their allotments at the end of round 4 (July 23 or thereabouts) to surrender their seats and fill in choices again! They allowed this again for the fifth round, called the spot round (not sure why). About 60% (maybe more: I dont have the actual stats)of the all India seats (the state seats were of course not surrendered as there was no chance of betterment) were surrendered and so for most seats the actual allocation took place only on the fifth (spot) round (August 2)! What a mess it was! With joint counselling, such crazy things will come on National TV (IITs will be involved) and no one will dare to try such things. You will be amazed how the JEE(mains) problems of 2013 and 2014 were completely ignored by the TVs and newspapers. 2013 was even worse (allotments were withdrawn time and again: NIT Patna got some 20,000 students allotted by mistake in one of the rounds). Had IITs been involved, the organising IIT's Director would have lost his job for sure.
So, you are right, there are dangers of delays in result declaration, but then major goof ups in JEE(Main) counselling will not happen (it is not CBSE, but a group of NITs who do the counselling). We need the IITs to lend their brand weight to the process!

Prashant Bhattacharji said...

so, the second stage (I guess that used to be called the 'main exam' in the early 2000s) is the one conducted by IIT - and that is an MCQ style paper?
I might be missing something here, but what's the point in having 2 complete MCQ tests.
MCQs - if done properly - can definitely test a lot. For example, in one Coursera class I took on Automata, proof related questions had been converted to an MCQ format rather successfully.
But I doubt they can successfully replace a complete test with narrative answers. Can't a bunch of narrative questions be thrown in?
Though obviously, I have no clue about how much of a headache it will mean for the administrators.

Dheeraj Sanghi said...

@Prashant, Many answers to your question. IITs insisted on a second test (JEE - Advanced) because they said that it will allow them to have a test with narrative answers. However, after JEE Advanced was agreed upon, IITs started having a cold feet about grading such questions in a consistent way, and what if RTI reveals some problems in grading. Frankly, the whole JEE issue within IITs is dictated more by fear of lawsuits than by the goal of admitting great students.

The second reason for a second test was that the first test was supposed to be like a public test where the result will be more uniform distribution, while the second test is supposed to be a test whose result has a long tail, so that we can select top 10,000 students, and it is less of a lottery. Of course, I don't think any faculty member in IIT system has ever tried to do a study of the question paper after the results are out and tried to figure out what were easy/moderate/difficult questions and what is more likely to result in a kind of distribution that we desire for our admission.

Third reason for the 2nd test was that we had no trust in anyone. Even if CBSE had agreed to let us decide the curriculum for the test, and also set the question paper, we would think that it is a sop for only one year, and the next year, all of them will force us to make changes. A 2nd test is immune from such pressures. Also, even though CBSE had conducted AIEEE pretty well till then, we at IITs felt that we do a better job of managing JEE, but we did not want to manage that for 10-15 lakh students.

Aryan Garg said...

One of the greatest flaws with the UG system of admission in IITs is the random allocation of Branches or streams regardless of whether the candidate has aptitude for the particular branch or not.

1. The first year should allow the student to see the real picture of the various streams.

2.the allocation of branches should be done after the first year of study.

siddharth jain said...

Sir just on reading some comment of yours regarding fear of RTI in doing grading of answer sheets, my very old wish of seeing my own IIT JEE 2004 answer copy of mathematics is reignited. Just want to know if you can help me in knowing can I get that answer sheet copy now and if to whom I should address this application ?

Dheeraj Sanghi said...

@Siddharth, The answerbooks are destroyed within six months of the exam.

jai said...

sir, can you tell when jee started?

Dheeraj Sanghi said...

@Jai, JEE started in 1962. It was called CEE - Combined Entrance Exam at that time, and included a paper on engineering drawing.