Quote:
Quote by: PatrickHenry  You could argue that. But without access to the methodology and data, you're just blowing smoke... |
On the contrary, without access to that stuff, I'd just reject their conclusion. Already from what I've seen concerning the poll, there seems to be enough bias to say their conclusion is unacceptable. I mean,
they're the ones making new claims and giving new figures.
Quote:
Quote by: Gods_Mercenary You have to be completely random and have a huge base, but even then, bias can be put into the wording in undetectable ways. |
Yeah, besides from the most obvious mistakes in the application of the statistical method itself, bias can still be present.
Quote:
Quote by: Chancellor And if the questioning was done face-to-face, even the inflection in the voice as the question is asked can change the results. It is exactly because of these kinds of things that I reject the entire notion of statistical sampling. |
Yes, this often occurs particularly in verbal polls. But sometimes when we're testing a drug, for instance, we can minimize the effects of confounding variables and bias by making the study/experiment blind, double blind, etc. Also, statistical sampling is used on non-human subjects as well, so
that form of bias you mentioned wouldn't really make much of an impact.
So, you shouldn't discard the entire field of statistics just because of the mistakes of numerous statisticians. However, they give a notedly bad impression.