With the NEET PG exam going on these days, a big issue is surfacing.
With a huge disparity between the different sessions, candidates are apprehensive as to how their overall ranking will be impacted.
We are consistently receiving E-mails from the candidates enquiring as to whether the disparity in the difficulty level of the different sessions will impact their overall performance and ranking.
To put the candidates’ confusion to rest, PrepLadder experts share in-depth details of the Normalization procedure used by the NBE in order to achieve standardization of scores.
Normalization Procedure – Details
Psychometry is the standard approach followed to achieve standardization of scores.
A Three - Step procedure is followed to standardize the scores:
Step I: Calculation of Raw Score
Raw score is calculated based on the candidate’s performance in the exam.
- The candidate is awarded +1 Mark for every ‘Correct’ answer
- The candidate is awarded 0 Mark for ever ‘Incorrect Answer/Not Attempted Question’
The significant aspect is the Total Raw score you obtain at this stage.
However, it is noteworthy that raw scores do not constitute Final Score/Percentile.
Step II: Equating Raw Scores
Next step is to equate raw marks.
The technique used to equate raw marks is ‘Item Response Theory’.
‘Item Response Theory’ implies a psychometrically supported statistical model used in the process of equating scores.
Item Response Theory takes into consideration the following Factors to equate marks:
- Performance of the Candidate
- Difficulty level of the MCQs
1. Performance of the Candidate
The general principle behind this is that ‘Candidates’ response to MCQ decides the Weight Age of the MCQ’.
‘Questions which are answered ‘Correctly’ by majority of candidates are considered as ‘Easy’ by the software and the software consequently assigns ‘Less Weight Age’ to such questions.
Likewise, Questions which are answered ‘incorrectly’ by Majority of candidates are marked as ‘Difficult’ by the software and the software consequently assigns ‘Higher Weight Age’ to such questions.
100 candidates attempted a question and 90% answered it correctly, then such questions will be counted as easy and will be consequently assigned lower weight age.
Likewise, if only 5-7% candidates answer a question correctly, software assigns higher weight age to such questions.
Also read: NEET PG: Overall Exam Analysis of First 3 Days (all 6 Sessions)
2. Difficulty Level of the MCQs
Difficulty level of the MCQs is ascertained using 2 factors:
a. Difficulty Index is calculated using the formula:
H + L
- Difficulty Index
- Discrimination Index
b. Discrimination Index is calculated using the formula:
H - L
H stands for
: Number of students with Correct Answer in ‘High Achievers’ Group
L stands for
: Number of students with Correct Answer in ‘Low Achievers’ Group
T stands for
: Total number of candidates who attempted that question
After marking the papers of a particular session, the papers are sorted in the descending order of marks and segregated into 3 groups.
The 3 Groups are
- Higher Achiever Group – Represented by the Upper 1/3rd candidates
- Low Achiever Group - Represented by the Lower 1/3rd candidates
- Middle 1/3rd - This group is excluded from Analysis. This group represents those candidates who attempt the exam through Guesswork and consequently answer some questions correctly and some wrong.
As a result, this group is not included while arriving at the analysis.
Step III: Equated Raw Scores are Scaled/Normalized
In order to normalize all the equated raw scores, a technique known as ‘Linear Transformation
’ is applied.
Linear Transformation comprises 2 techniques:
- Normal Distribution
- Standard Normal Distribution
These techniques are applied to arrive at the Z score.
Z – Score is given by the formula
How Z-Score Varies in Easy & Difficult Sessions
For Easy Session
Assume that 100 candidates appeared in an easy session.
It is quite evident that the Mean score will be high for an easy session.
Assume that the Mean score is 70 and Standard Deviation is 15.
Suppose a candidate has secured 85 marks as the raw score.
Then, Z-score = (85 - 70)/15 = 1
Z-score in this illustrated easy session is 1.
For Difficult Session
The Mean Score will obviously be Low in a Difficult session.
Assume that mean score is 45 and a candidate has obtained 75 marks.
Assuming standard deviation to be 20 in this session,
Z score = (75 - 45)/20 = 1.5
In a difficult session, Z- score is 1.5
By the illustrated example, it is clear that Z –score is higher in the Difficult session.
How Z-Score impacts Final Normalized Score/Percentile
Now, we will see that what impact Z-score has on the final percentile/Normalized score.
Final Percentile or Normalized score is computed using the formula
Target Mean + (Z-score * Target Standard Deviation)
Here, it is noteworthy that Target Mean & Target Standard Deviation are same for all sessions.
The Methods to Compute Target Mean & Target Standard Deviation is as follows
They are computed using either of the 2 methods:
Mean Score is calculated for each session separately and then the Mean values of all the Sessions are used to compute the Target Mean & Target Standard Deviation
For each session, Marks of Top 1% or Top 10% candidates are utilized to compute the Target Mean & Target Standard Deviation.
Now, it is clearly evident that Z-score is the sole factor which impacts your Overall Percentile and Ranking.
Finally, all the computed Normalized scores are arranged in the Descending order to arrive at the Final Merit list.
It is clearly evident now that disparity in different sessions will not impact your Overall ranking.
Normalization procedure used by the NBE will standardize the scores of all the candidates without any bias.
Therefore, you must not be worried in case your session was difficult or must not think that candidates appearing in the ‘Easy’ session will benefit.
We hope that this post will definitely help you understand the Normalization procedure used by NBE.
Stay Tuned for More Updates!!