A very common question that bothers many CAT-takers is the scoring
methodology used to assess their performance. Specifically, the mystery
around normalization makes students wonder about its impact on their
final scores. In this article, let us try to decipher the normalization
process.
First of all, CAT awards +3 marks for correct answers and -1 mark for incorrect answers. There is no penalty for questions that have not been attempted. The scores obtained by this method are then subjected to the normalization process. The scaled scores that are obtained through this are then converted to percentiles. The percentiles are then used for the purpose of shortlisting. Further, a CAT-taker’s scorecard will show the candidate’s sectional and overall scaled scores as well as percentiles.
The CAT notification says that “The process of Normalization is an established practice for comparing candidate scores across multiple Forms and is similar to those being adopted in other large educational selection tests conducted in India such as Graduate Aptitude Test in Engineering (GATE).”
So, let us take a look at the GATE normalization process. Ignoring the actual mathematical workings (which are cumbersome and would not really add to our understanding), the GATE process assumes that the distribution of abilities of candidates across test slots/sessions is the same. It then looks at the mean and standard deviation of scores across slots in order to compare the difficulty level of the test in each slot.
CAT 2014 had 4 slots and about 1.7 lakh test-takers. Thus, each slot had close to 42,000-43,000 test-takers. Further, slots were assigned to candidates randomly – test-takers could not select their own slot. Thus, it is quite reasonable to assume that the distribution of abilities of candidates across slots would be comparable (due to the large number of test-takers in each slot and due to the random allocation of slots to test-takers). Further, with just 4 slots this time, CAT 2014 is not likely to have wide variation in difficulty levels.
Hence, to sum up, normalization should not cause any problems for you. Your actual CAT performance should reflect in your scaled scores this year.
All the Best for your results!