Normalization in NEET PG 2024 Discussion Queries Part II

[@Tenshin-pz2kn]

Hope paper 2 topper doesn't score more than or equal to paper 1 topper. Because, then it will not fo justice for the rest of paper 2

[@basavanandv8839]

Exactly my query...

[@Ramya12345-t]

How much rank can differ from actual rank if we are in shift 1??

[@DrNRSwathiSai]

Some of them are saying in normalisation process marks will be added.. but u r saying merging of scores.. which is reliable

[@ZyNerdNEETPG]

This the process mam. NBE released this process a day before the exams. Till the process of ranking this is how it will be. your ranks will be defined this way. however post that in order to put up marks - one set may be given equivalent marks like another shift. that is a separate process.

[@DrNRSwathiSai]

@@ZyNerdNEETPG this mean, our raw score remain the same.. for example I score 460 marks in tough section, last year that marks gets around 40k rank, with this normalisation can i expect better rank..

[@Lapakchu]

​​@@DrNRSwathiSaidid u failed math or didn't take during 10+2 exam! A simple percentile based exam does not depend on exam levels. Pathetic medico

[@vijaysekhar7719]

Sir, why the NBE is not conducted in one shift only this year . what is the reason behind this.

[@ZyNerdNEETPG]

yes sir. it would all have been better if they had conducted in one shift even if exam was delayed. this has created so much of problems. the mentioned reason is inability of getting enough centers. they could have held it a little delayed.

[@Nema.Sushil]

Can you please check the Normalization of the two sets ( 2022 and 2023) which you have quoted using the formula used by SSC which uses mean and standard deviation and the marks scored by top 0.1 percent of the students in each shift. I am not able to copy past that formula here. I have used that formula on the merit list of MPPG 2023 ( based on NEETPG 2023) and I found that the variation is of only 1 rank using this formula whereas if I used the percentile based formula as per the erstwhile AIIMS and quoted by NBE then the RAN difference was going up to 60. What i did was that I took the top 2500 ranks of MP merit list. Sorted that merit list alphabetically and then using the Excel Randbetween formula divided them in two sets(shifts). Then using there actual marks I calculated the merit list using both the method. I found the SSC method better as it was resulting in rank deviation of only 1 whereas the percentile formula was giving a rank difference of up to 60 when compared with their original rank. It will be good if you also use the two formula on your two data sets and see whether you are getting the same observation as mine. If the observation is same then we would know the better formula but yes whether we can convince NBE about that or not is a different story but we can put that in public domain for a better decision if possible.

[@ZyNerdNEETPG]

we will check this sir. we had done this with a sample of 50% taken from 2023 and 50% taken from 2022 merging both assuming that as two different shifts. the difference is that 2022 had lesser candidates(1.8 lakhs around) but still we went ahead considering what if so. the actual difference in numbers per shift in 2024 might be similar. the rank difference of 50-60 might also be due to clustering at a few points which normalization slightly misses. we can check. will try the SSC method and let you know. sir.. we didnot use the randbetween but used alternate ranks so that the distribution is exactly maintained. both random and this method should have major variations sir.

[@Nema.Sushil]

@@ZyNerdNEETPG thanks

[@pandharinathsamant4832]

How is difficulty levels in both shifts addressed,? Percentiles is on number of candidates who attended.

[@ZyNerdNEETPG]

@@pandharinathsamant4832 after percentiles are shared you can see both toppers (one with 725 in shift 1 and another with 705 in shift 2both are rank 1 tied). 705 even though got a difficult paper got rank 1 even after scoring lower due to difficult paper. . same happens for those in shift 2. that's the adjustment sir.

[@kiransrinivas7179]

Sir can you make a vedio of normalisation of previous batch like what you said ,comparing 2022 and 2023 result

[@ZyNerdNEETPG]

will try sir. not sure if it would be going overboard ? we thought we would explain this in detail and already a lot of negative comments.

[@pandharinathsamant4832]

Where is difficulty level addressed in normalization?

[@drmitulsinh]

Sir i able to solved tough question in evening sessions, will it affects my net result

[@ZyNerdNEETPG]

if you have done well in a tough question in the evening session, then your rank will be higher (better) in the tough session. that means it will reflect better in the overall rank too sir.

[@vanditachoudhary]

If I performed better in difficult questions in shift one then it will be better for me or not sir?

[@bhoopathybru7347]

Equivalent score for state counselling should be based on new rank/new percentile obtained after normalisation process

[@ZyNerdNEETPG]

it will be based on the new percentiles mapped sir. a raw equivalent score will be taken (with anyone of the shift scores as the base). however which shift will be taken as based - we are not sure. this impact is only for states where inservice 102030 % is applicable

[@amlangoodsongradhamitra8250]

sir what can be rank at 480

[@pranjaltoshniwal1999]

⭐⭐⭐⭐⭐