- Nov 14, 2012
- 12,164
- 19,684
- AFL Club
- Carlton
There is no transparency in terms of polling.
How is it even done? I don't understand how you can get a accurate representation of public opinion when you are polling truly random people
Their disclosed method is to weight responses, (presumably by age, education, occupation, gender, ethnicity etc) to arrive at a representative sample of the polity.
Nothing wrong with it in theory except it potentially magnifies outliers and could be open to manipulation. This seems especially plausible when you reduce your sample from 2000 to 1600.