PISA scores and missing students

 

While I was working very slowly through some of the technical appendices to find how PISA estimates “educational equity” (which turns out to be mean that there is a lack of educational equity anywhere that there is a strong impact of SES on educational performance) Steve Sailer had moved fast into Annex 2a, from which he was able to estimate the representativeness of the samples tested. It looks as if representativeness was often pretty low, which gives some latitude if you decide to get your best students into the exam room, and let the weaker ones be classified as “missing” for the exam day.

http://isteve.blogspot.co.uk/2013/12/pisa-which-countries-not-to-trust.html

 

This looks like being a very important finding. I had spoken vaguely about the possibility of nations “gaming” the test by teaching to the test items, but keeping a good proportion of the weaker students away is a much simpler strategy.