RSS 2.0 Feed

» Welcome Guest Log In :: Register

Pages: (2) < [1] 2 >   
  Topic: Design Prediction - Take Two< Next Oldest | Next Newest >  

Posts: 37
Joined: June 2004

(Permalink) Posted: April 14 2005,12:21   

Quote (Michael Finley @ April 14 2005,10:46)
Consider the following example:

In the U.S. around 90% of blacks vote Democratic. Suppose I meet a black voter, and all that I know about him is that he's black and he's a voter. Can I reasonably predict that he votes Democratic, i.e., is it probable that he votes Democratic?

I think I see where your argument is going.

While this response may seem like a nitpick it does illustrate an important point.

If we take it literally that ALL you know about this black person is that he or she is a voter and black - the person in question is randomly selected from every black person in the world who has cast a vote - the answer is "no".  Because there are many black voters in this world outside of the U.S. and they, of course, vote for the governments of their own nations.

The important point is that you cannot extrapolate from a sample unless you know that sample to be representative.  The U.S. population of black voters is not representative of the world population so your "prediction" is a product of poor sampling methodology.

  32 replies since April 14 2005,10:46 < Next Oldest | Next Newest >  

Pages: (2) < [1] 2 >   

Track this topic Email this topic Print this topic

[ Read the Board Rules ] | [Useful Links] | [Evolving Designs]