Now we can say beyond doubt that in this forum users like
@Pooja_11 and others are have real contacts inside UPSC. So this user Pooja had given some info on the new moderation formula which UPSC has started to follow from this year. I am pasting what she told about that here. Let us analyze the implications of this new formula in the optional selection, performance. If anyone knows the moderation formula followed before this one please comment on that also.
"I again talked tob> my insider..he told the moderation formula this year has been devised in such a way that the number of candidates invited for interview from each subject will be approximately in proportion to the number of students in each subject in mains.this has been devised after very much thinking given to it..this is to eliminate any type of bias of which allegation are there agains them.they are hoping that it will provide more levrl playing field...it is explained as follows
1.suppose there are 15000 students appearing in mains..And total candidates invited for interview are 3000..
2.suppose sociology has 3000 students in mains..it is 20% of total mains candidates that is 15000 ..now they will try to do moderation in such a way that approximately 20% students who are invited for intrview are from sociology..so approximately 600(20% of 3000) students are invited from socio..
3.suppose maths has 450 students appearing in mains.it is 3% of all the students appearing in mains (15000)..so they will do moderation in such a way that approximately 3% of all the students who are invited for interview are from maths..so they will invite almost 90 (3% of 3000) students in maths..
4.this also applies to other subjects..
Again this is not the exact formula..whatever i could understood from the talk on mobile, i have reproduced..After reading this formula i have the opinion that it is better to scrap this optional shit altogether! See, literally UPSC is eliminating a candidate choosing one optional even though his answers are far more superior than an average guy who is taking some other optional. Yeah competitive exams are all about relative performance but checking relative with in relative is bullshitting and nothing else.
IMO, Moderation formula will also include factors like adding more regional language students, literature students some other bullshits. Taking all those shits, if we compare the raw marks gotten and moderated marks it would be a disaster for many candidates and it would turn heavenly surprise for some others.
Comments
Still with no transparency and questionable reliability, is it wise to take such an important decision based on hearsay?
Anyone has any data ( Mrunal? ) on optional wise number of candidates appearing for Mains 2013 and final list 2013?
Indeed, removing optional was a more logical step rather than removing CSAT , ( do not attack me, everyone in entitled to his opinion ). Can someone here get this message to the powers that be? ( ABVP members? )
I will give some data from the file.
Economics had 38 final selections (~3% of total~1240) while those appeared for Eco mains were 183 (~1.2%), implying a 20% conversion from mains to final result. 2012 nos. were quite similar with 50 final selections and 250 appearing for mains.
The same thing played out with Law (2013: 34/140, 25% success rate); Law and Eco being two most successful conversions from Mains to final selection.
The overall selection was higher (at 177) for Geography and 176 for pub ad. but candidates appeared for these two were at 3115 and 2802 respectively.
PS: While I agree that weird things have been done by UPSC earlier as well but the new moderation theory as being propounded is not statistically sound. Having the same average marks for optionals is a better improvement though not perfect. Having the same average for optionals taking only middle 90% (i.e. removing top and bottom 5% outliers) might be a further improvement.