The year 2024 has been a contested year for examinations in India, with many aspirants raising questions and taking to the streets about various alleged discrepancies in them.
One of these issues that was contested by students was the normalisation of scores in some of these exams, particularly in the National Eligibility-cum-Entrance Test Postgraduate (NEET-PG) and the Uttar Pradesh Public Service Commission (UPPSC) exams.
To recall, the NEET-PG 2024 was conducted on August 11 in two shifts, and the National Board of Examinations in Medical Education (NBEMS) normalised the scores from both shifts utilising the formula used by the All India Institute of Medical Sciences (AIIMS) to normalise scores in its entrance exam.
After the exam, aspirants petitioned the Supreme Court of India, alleging a lack of transparency in various aspects of the exam, including the normalisation process used to calculate the scores. The matter is subjudice.
Similarly, the UPPSC announced on November 5 that the exams for the Combined State/Senior Subordinate Services (Prelims), commonly called PCS (Pre) and Review Officer/Assistant Review Officer recruitment would take place in two shifts, and the scores of the candidates would be determined through normalisation.
This led to UPPSC aspirants launching a four-day long agitation in front of the commission’s headquarters in Prayagraj, Uttar Pradesh, demanding that the proposal be taken back and the exam be conducted in one shift.
Following the agitation, the UP Government announced on November 14 that the exam would be conducted in one shift.
These instances beg the question: What is meant by the normalisation of scores? How is it done?
The purpose of normalisation
Normalisation in statistics is a method used to convert values measured on two or more different scales to a common scale.
In the context of examinations conducted in multiple shifts, normalisation is used to ensure fair assessments of the candidates’ performances across shifts, by using similar exam settings.
“When exams are conducted in multiple shifts, different sets of question papers are used. The exams are also conducted at different times of the day. All these variables impact the difficulty levels of the exam, and eventually the distribution of marks significantly across the two shifts,” explains Prof Tathagatha Sengupta, a Professor at the Homi Bhabha Centre for Science Education, Mumbai.
In such a situation, he says that it would be unfair to ignore these variables and club the marks across both shifts to determine the final ranks.
“At this stage, it is important to bring the marks obtained by candidates in both shifts to the same scale and equate them. Hence, the scores are normalised,” he says.
How are the scores normalised?
The methods and formula used for the normalisation of scores in both the NEET-PG 2024 and UPPSC exams are fairly similar.
Here is the process:
Step 1:
The raw scores of the candidates, ie the scores obtained by the candidates in the exams are calculated and collected.
Step 2:
The candidates’ percentile scores are calculated from their raw scores. The percentile score of a candidate indicates the percentage of candidates who scored equal to, or less than them.
The percentile scores are calculated using the following formula:
P = (Number of candidates who scored below X / Total number of candidates in a shift) ✖ 100
Where P is the individual percentile score and X is the candidate’s raw score.
For exams with multiple subjects, both the total percentile and subject-wise percentiles are calculated.
As per the AIIMS normalisation procedure, the percentile is calculated up to seven decimal places, and the percentile for the UPPSC would have been calculated up to six decimal places.
Step 3:
After the individual percentiles are calculated, the merit list is created.
First, the percentiles from both shifts are merged to create a single list, and ties are resolved.
In exams with multiple subjects, the ties are resolved based on the percentiles in each subject.
For example, if two candidates have the same total percentiles, whoever gets the higher rank is determined by whoever has the higher percentile in Subject 1. If Subject 1 percentiles are unable to act as tie breakers, Subject 2 percentile is considered — and so on, till the last subject.
For exams with a single subject, the older candidate would be ranked higher.
Based on these tie-breakers, the final merit list is formed for both shifts combined, with the candidates’ percentiles.
