The Real Indian IQ

[G.Subramaniam]

Errors in PISA 2009 Indian scores - part 3

-----------



http://pisa2009.acer.edu.au/interactive.php



is an interactive website to look at how PISA score changes under various factors



--



My first concern is whether PISA over sampled linguistic minorities deliberately to lower the score



In HP, the medium of instruction is Hindi and the linguistic minorities are Tibetans and Punjabis



In TN, the medium of instruction is Tamil the main linguistic minority is Telegu



--



In the interactive PISA site,



under School variables, SC08Q01 asks 'Is the test language same as the first language of the child'



This would catch linguistic minorities, for example a Telegu child in a Tamil classroom, answering a PISA exam set in Tamil



or



A Tibetan or Punjabi child in a Hindi classroom, answering a PISA exam set in Hindi





Category-1 means over 60% of the students are from a linguistic minority and Category-6 means all students are from the linguistic majority



--



In Tamil Nadu, 75% of the student sample was in Category-1 and they scored 344 mean for PISA math, and

only 20% of the student sample was in Category-6 and they scored 378.



In any good sampling system, 95% should be in category-6 and just 5% of student sample in Category 1



The PISA mean was 351. But for Tamil students answering PISA in Tamil, the score was 378

The gerrymandering artificially reduced score by 27



At 378, the math score is same as Rajasthan in TIMMS



--



In HP, 83% of the student sample was in Category-1 and they scored 327 mean for PISA math

and only 10% of the student sample was in Category-6 and they scored 401



In any good sampling system, 95% should be in category-6 and just 5% of student sample in Category 1



The PISA mean was 338, but for Hindi students answering PISA in Hindi, the score was 401

The gerrymandering artificially reduced score by 63



At 401, the math scores equal Orissa in TIMMS