Monday, November 21, 2011

Stats for JEE-2011: Income Effect


Just skimmed through the report on JEE-2011. It has all kinds of sociological data (though not fine-grained enough for us to perform our own analysis). Here's an interesting set of data on the candidates' parental income.

According to parental income

Table A13 shows the zone wise distribution of candidates according to annual income of parents. Out of the 468280 candidates who appeared, 146882 (31.36%) had a parental income of less than 1 Lakh. 195220 candidates (41.69%) had a parental income between 1-3 Lakhs. Another 19.3% have a parental income between 3-6 Lakhs, while 4.87% have parental income between 6-10 Lakhs. 12520 candidates (2.67%) have parental income exceeding 10 Lakhs. The percentages for qualified candidates for the five slabs of parental income are 17.1%, 31.7%, 30.47%, 11.86% and 8.86% respectively. The income slabs used in JEE 2010 data analysis was more or less similar. However it is observed that a large percentage of appeared as well as qualified candidates belong to the low to middle income groups.

The report doesn't tease out the implications on the income effect on JEE outcomes, but we certainly can. When we do that, we find a clear trend:

Income No. of JEE Takers No. of JEE Qualifieds Pass %
Less than 1 lakh 146882 2258 1.54
1-3 lakhs 195220 4183 2.14
3-6 lakhs 90560 4021 4.44
6-10 lakhs 22835 1565 6.85
More than 10 lakhs 12520 1169 9.34
Total 468280 13196 2.82

In other words, a candidate from the top income slab is 6 times more likely to get through JEE than one from the lowest slab. This advantage is "only" 4 times when the comparison group is from the second lowest slab (which also has the largest number of candidates).

16 Comments:

  1. inthearmchair said...

    The next logical question is why. Of course wealthier candidates are able to prepare themselves better -- a process that probably begins in very early childhood.

    But I wonder if there's also a self-selection effect. E.g., a son of a wealthy businessman might have no interest in killing himself to get into the IIT unless he feels he has a good chance, while a son of a poorer businessman might not feel he has any choice but to try.

  2. Ankur Kulkarni said...

    What would be interesting is the trend over time. My guess is that these fractions would have shown a trend where over time (especially since the time coaching classes have boomed) more students from a lower income bracket have been able to pass the JEE.

  3. Desi Babu said...

    Good Analysis Prof. Abinandan!

    One thing most people tend to forget or neglect is the (slightly controversial) "gene-pool" effect. Higher income parents are also
    associated with better educational backgrounds, and sometimes, they can pass on certain genetic traits, that help the kids make the cut.

    This is also clearly visible in "rich" school districts in the United States, but again, for political correctness, this is simply labeled as a "rich" school effect, i.e, better facilities, teachers' salaries etc. If it was only money, then Beverley Hills 90210 would be churning out Einsteins by the dozens, wouldn't it?

    Peace!

  4. WebMiner said...

    @desi -- A positive correlation is expected, but the important issue is the location of most people wrt the knee of the wealth-quality payoff curve. In the US the quality knee is achieved by most schools, not just 90210. (In fact, 90210 has fashion and other distractions.) Indian parents making less than, I don't know, 5 lakhs per year now probably cannot afford some facets of quality that seriously matter to their children. An important hidden variable here is the locality in which the kids grow up, and the intellectual and professional quality of neighbors and their kids. Nothing affects a kid more than the career moves of their peer group. In many large Indian cities you have to shell out over a crore to get housing in localities that foster the right kind of peer group and competition to get through JEE. The positive correlation should surprise no one. Of more interest is where the knee is located.

  5. Debajyoti Datta said...

    @inthearmchair Self selection is an interesting point, however starting from there I can argue that candidates from poorer families would have a greater incentive in clearing the exam while those from richer families know they have a safety net.

    @Desi Babu It will be impossible to untangle the effects of socioeconomic factors from genetic factor from this data. However, the assertion that genetic factors may explain this disproportionate finding is at best questionable and at worst misleading.

    Intelligence is ofcourse influenced by genetics but it is also heavily influenced by socioeconomic status. Infact low socioeconomic environment is the dominant factor influencing intelligence in impoverished families whereas genetic factors explain the variability in well to do families. This means that if one is born poor, one can never realize the full genetic potential for intelligence. See the works of Turkheimer et al.

    It will be more evident if the data from private colleges are available.

  6. Digbijoy Nath said...

    I was wondering if poverty is a reason for one's talent to be restricted in blooming, how was Abel - one of the greatest mathematicians ever who died at a young age of 26 or something like that - lived in extreme penury and still accomllished work to keep mathematicians busy for the next 500 years.

    Great ones like Ramanujam, A P J Abdul Kalam, Faraday, Marie Curie & the list is endless - all were, in their contemporary times, from lower middle class families. Marie Curie, wrote her daughter Irene Curie in the former's biography, had to live with one jacket throughout her PhD life.

    Getting into IIT for an undergraduate degree, I strongly believe, does not define success or talent. In fact, it does not define anything. Life's true challenge and journey starts from there - and not ends there unlike what many common people think. There was this guy securing all India rank 1st in general category, in a certain year, like 10 years back, (I won't name him), who had failed in his first year of undergrad in the IIT he took admitted. Nobel laureate Dr. Venkatraman could not clear IIT JEE. There are uncountable great scientists doing front end research in the USA, who did undergraduate in ordinary engineerin colleges in India.

  7. Debajyoti Datta said...

    @Digbijoy Nath That the great ones succeeded despite adversity is a testament to their greatness but surely it would have been better if they hadn't had to face such adversity. For example, Ramanujam died at a young age from tuberculosis, a disease common in the poor.

  8. Prasad14 said...

    Such analysis definitely echoes SOCIAL MOBILITY in broad context.

    In the western standard, more than half of Indian population would belongs to 'less-privileged socio-economically' to contribute equally as creative western world does either for scientific innovations or for modern-cultural drive(for eg. super-duper 3D science fiction films from Hollywood or facebook).

    Even in developed country like in UK, the government is adapting to a policy to close opportunity gap for 'less-privileged'(term relative to UK government standard)

    some of them here,

    http://news.bbc.co.uk/2/hi/8160052.stm

    (includes "report on fair access to the Professions")

    http://www.bbc.co.uk/news/uk-12339401

    http://www.guardian.co.uk/society/2011/sep/23/upward-mobility-women-low-earners

  9. Ankur Kulkarni said...

    @Digbijoy Are there examples of any one from a poor background succeeding at a highly competitive (rat) race?

    These stats actually prove that talent is not the only thing that count in the JEE - for talent ought be uniformly distributed across all levels of parent's income.

  10. MP said...

    I have a hard time understanding what JEE is trying to do. It could be testing for:

    1)High Intelligence
    2)Motivation
    3)Hard work
    4)Insight

    The popular idea is that it tests for 1. But the cram schools show that 2 & 3 in combination can create results approximating 1.

    I personally think it should be testing for 4. But no one knows how to do it. Intelligence is commonplace, insight is rare.

    It should also test for 2, because 2 can usually trigger 3, but once again no one knows how to do it.

    Also, 4 usually comes with 1, but 1 does not guarantee 4. Neither 4 or 1 guarantees 2 & 3; in fact, they might exclude 2 & 3.

    I also doubt whether any of the test designers have sat and thought about it in these terms. Including a good educational psychologist in the design process would be a good idea, if you can find one such person in India!

  11. WebMiner said...

    There may be some "undisclosed wealth effect" here. Tax rates are highly sensitive to the 2--8 lakh/y range, and JEE data belongs to sarkar, so I am sure some richer parents declared less than their true income, reducing the success rate in lower income brackets. Meanwhile, the ones who did (could) not fudge their income in the highest brackets are salaried folks with very high qualifications, which predispose their children toward success.

  12. Raj said...

    I find it very hard to believe this data which says 2200+ of the qualified candidates are from what is practically below the poverty line. I think people generally have under reported the income (especially those that are not salaried employees) in the application. The exam that has probably created the most 'cram school' business in the world gets 15% of its student from families that can barely afford basic needs of life? Come on! The reality is probably more depressing than this (i.e. poorer students have far lower chance than one in six of getting into IIT). This report probably is not at fault here. It is the people providing that data that have screwed it up.
    Another interesting stat is 37% of the qualified candidate have said that they prepared for the exam through 'self study'. Really?

  13. Dheeraj Sanghi said...

    @Webminer, @Raj, I don't believe JEE data either. People under-report like anything. But I have no reason to believe that some groups do it more than the others. So, in reality, just the numbers on the slabs will be different, but higher slabs are likely to still have higher rate of selections.

    @Ankur, with organized coaching such as Kota becoming available, a larger number of students from small towns and semi-urban places have started appearing in JEE and succeeding in it. It is now possible for them to go to Kota or Hyderabad or for that matter Kanpur and stay in a "hostel" and get coaching. But they are, by no means, poor people. Since everyone goes for coaching, the competition has become so intense that self-study cannot get you a seat in IIT (despite what JEE forms self-declarations say). And coaching is expensive. Poor cannot afford it. The ratio of poor people in IITs have come down - at least this is what I feel when I interact with students today, including students from remote areas.

  14. Virus said...

    Nice analysis.
    It will be great to do a similar analysis for each category of applicants (open and reserved categories). That may provide some insights on how effective the system of classification is in reaching out to a wide range of students. I do not intend to present any opinions here but just interested in how the system may be made more effective.

  15. Poornima said...

    I think it should be stressed that a candidate of the highest income bracket is also twelve times less likely to appear for the test and six times more likely to pass it than the lowest income bracket.

  16. Soundarya said...

    higher income parents were at some point in the past born from lower income parents. and their children could have got into iits through access to coaching classes. hence although the family may be only literate a few generations back, a child may have got into iit. thus the correlation that high income parents with better educational backgrounds have better genes is a spurious argument (then even their ancestors would've had good genes, but why were they poor?). there is no evidence in biology in my knowledge to establish that better income and education is genetically determined.