Can statistics be manipulated to support both true and false arguments?

[Loren Booda]

Would you agree that statistics can be mathematically misrepresented to prove most related arguments both true and false?

 

[CEL]

Would you agree that statistics can be mathematically misrepresented to prove most related arguments both true and false?

Data mining allows any statistical result.

 

[russ_watters]

Depends on what you mean by misrepresented. You can certainly prove whatever you want with lies, but I don't think that's what you mean...

 

[FredGarvin]

I would say that proper statistical models usually rely on a very strict application of techniques and assumptions. It is an area that is not well known by most people so it is very easy to use them improperly. It is no different than any other form of mathematics. People misrepresent statistics. The math does not lie.

 

[Loren Booda]

I would say that proper statistical models usually rely on a very strict application of techniques and assumptions. It is an area that is not well known by most people so it is very easy to use them improperly. It is no different than any other form of mathematics. People misrepresent statistics. The math does not lie.

That's what I needed to hear.

 

[lawtonfogle]

Would you agree that statistics can be mathematically misrepresented to prove most related arguments both true and false?

If you mean use a statistic in such a way which is dishonest/incorrect, but which the average, non-statistically minded person, would accept as a correct use of the statistic, then I would agree it can often be used this way. I do not know if one could say 'always'.