A blog writeup on CBSE, ISC versus state boards and how the difference in the quality of schooling makes simple percentile comparisons problematic.
Also, see the comments (click on the comments link at end of the blog post mentioned in 1. above) whether normalization is such a simple, straightforward process or needs far more study and thought.
The data compiled by MHRD (2008 - so slightly dated) shows huge variations in the sizes of Boards. The smallest has only 325 students and several state Boards have a few thousand students. In contrast the largest Board (UP Board) has over 16 Lakhs students. The pass percentages vary widely from 48.24% to 98.15% - (see the data for higher secondary i.e. class 12 not high school which is class 10). Can we apply simple percentile based statistical normalization techniques to such numbers?
Even within a single Board (CBSE) the pass percentages vary significantly from one region to another. This data for 2008-2010 is from the CBSE website. Notice the huge difference (over 20%) between Chennai and Guwahati regions.
Board results can vary by large amounts year-to-year. For example this year (2012) news reports say the pass percentage has gone up by almost 10% in the UP Board. In 2008 there was a huge fall in pass percentage when 'self-centres' for the Board exam were abolished after the government in UP changed. Such changes make percentile based comparisons across years meaningless.
Some Boards (CBSE) already practice a moderation policy to ensure that pass percentages across years stay close to each other. Is normalization across Boards with different (or no) moderation policies a fair process?
The assumptions made by the expert committee on normalization from ISI, Kolkata on normalization probably do not hold.
The original ISI report of Aug. 2011 says very clearly that normalization across Boards is not possible.
This HT article (2nd Jul) clearly brings out that normalization is not scientifically supported. One of the authors of the report says "We were then told very clearly that we were not to worry about whether scores could be compared, but to only provide the best way to compare them".
This whole sorry episode has not ended. The NIT Council has gone ahead and endorsed inclusion of 40% Board percentile marks in the ranking of students for admission to NITs.