11 November 2016

polls

polls are like the best possible guess, better than blind guess. reality is another thing. like if i say you have 90% chance of winning it doesn't mean you will win.

e.g. if i interview 1,000 people out of 10,000 by random sampling and if all 1,000 people happen to say they love apples, it doesn't follow that all 10,000 people must love apples.
...


some pundits say that another reason is that people lie, but i don't know why they have to do this. 


mine is a more philosophical one. of course some may argue that if random sampling is done better the probability should be so reflected. but then in reality there is no perfect random sampling. from an empiricism point of view, as opposed to a rationalist point of view, the probability theory cannot be proved in the real world as we don't do things perfectly and we don't master eternity.

No comments: