02-25-2012, 11:17 AM
(This post was last modified: 02-25-2012, 11:40 AM by G.Subramaniam.)
Errors in Indian PISA score - Part 4
---
The next area where gerrymandering has occurred is in over-weighting rural areas and in the urban areas, deliberately selecting bad schools
--
Among the school variables, the relevant factor is 'school community'
and is represented by question = SC04Q01
--
Category-1 = Village
Category-2 = Small Town
Category-3 = Town
Category-4 = City
Category-5 = Large City
--
I would normally expect the score to go up by 10, at each step, A 10 score is about 1.5 IQ
The reason is that as you get more urbanised, the schools get better, healthcare gets better, the percentage of upper-caste, higher IQ segment increases and other similar factors
--
In Himachal Pradesh,
67% of the sample is Category-1 or village with a mean math score of 329
15% of the sample is Category-2 or small town, with a mean score of 379
18% of the sample is Category-3 or town, with a mean score of 339
0% is Category-4
0% is Category-5
I smell a rat here.
First they have given 67% weighting to village and 0% weight to both City and Large City
Instead of seeing a rise of 10 points as we go from Category-2 to Category-3,
There is a drop of 40 points going from Small Town to Town
They must have searched for the worst slum school in Town
---
In Tamil Nadu,
44% is Village with mean math score of 332
9% is Small Town with score of 337 ( just a gain of 5 is fishy )
28% is Town with mean score of 366
12% is City with mean score of 386
7% is Large City with mean score of 376 ( A fishy loss of 10 instead of gain of 10 at this step )
Tamil Nadu is a highly urbanised state with Village and Small Town combined at about 25% of population
whereas this survey has given them 53% weightage
Also in TN, brahmins are heavily urbanised and most are in the large City category , which has been under weighted
--
In both states, with proper rural-urban weightages, I would expect scores to rise by about 20 or 3 IQ
---
The next area where gerrymandering has occurred is in over-weighting rural areas and in the urban areas, deliberately selecting bad schools
--
Among the school variables, the relevant factor is 'school community'
and is represented by question = SC04Q01
--
Category-1 = Village
Category-2 = Small Town
Category-3 = Town
Category-4 = City
Category-5 = Large City
--
I would normally expect the score to go up by 10, at each step, A 10 score is about 1.5 IQ
The reason is that as you get more urbanised, the schools get better, healthcare gets better, the percentage of upper-caste, higher IQ segment increases and other similar factors
--
In Himachal Pradesh,
67% of the sample is Category-1 or village with a mean math score of 329
15% of the sample is Category-2 or small town, with a mean score of 379
18% of the sample is Category-3 or town, with a mean score of 339
0% is Category-4
0% is Category-5
I smell a rat here.
First they have given 67% weighting to village and 0% weight to both City and Large City
Instead of seeing a rise of 10 points as we go from Category-2 to Category-3,
There is a drop of 40 points going from Small Town to Town
They must have searched for the worst slum school in Town
---
In Tamil Nadu,
44% is Village with mean math score of 332
9% is Small Town with score of 337 ( just a gain of 5 is fishy )
28% is Town with mean score of 366
12% is City with mean score of 386
7% is Large City with mean score of 376 ( A fishy loss of 10 instead of gain of 10 at this step )
Tamil Nadu is a highly urbanised state with Village and Small Town combined at about 25% of population
whereas this survey has given them 53% weightage
Also in TN, brahmins are heavily urbanised and most are in the large City category , which has been under weighted
--
In both states, with proper rural-urban weightages, I would expect scores to rise by about 20 or 3 IQ