Tim Morten
Well-known member
- Joined
- Apr 27, 2008
- Messages
- 1,031
- Reaction score
- 0
- Points
- 0
Oh, I understand the disconnect on the stats now. Terry and Daniel are thinking about Gallup polls and the like, which are used to decide nothing due to their insufficient samples. I was thinking about the criteria actually used to elect / make decisions, which of course require statistically significant samples. Some Gallup polls align with the actual result, some do not. No actual election would be conducted with such a small sample percentage. Interestingly in the US, the Electoral College is the body primarily sampled when we make decisions since we're an indirect democracy; the requirement for actual voter turnout is quite low.
Anyhow, the quoted 'margins for error' are indeed math, but not actually objectively predictive measures, just projections that a statistician would categorize as speculative in the absence of a statistically significant sample... often 5% of the total population is the minimum criteria, though some might argue for 2%. There is a substantial chance of drawing a false conclusion by taking a sample size smaller than that.
I'm sure the response will be that somehow the miniscule sample we have should magically be interpreted as definitive, since it aligns with the opinion that some here want to champion. Just happy to understand the source of the disconnect.
Anyhow, the quoted 'margins for error' are indeed math, but not actually objectively predictive measures, just projections that a statistician would categorize as speculative in the absence of a statistically significant sample... often 5% of the total population is the minimum criteria, though some might argue for 2%. There is a substantial chance of drawing a false conclusion by taking a sample size smaller than that.
I'm sure the response will be that somehow the miniscule sample we have should magically be interpreted as definitive, since it aligns with the opinion that some here want to champion. Just happy to understand the source of the disconnect.