Statistics are not good enough for the naysayers?

[RaulVB]

Ask US Supreme Court Justice Stephen Breyer.

The "highest tribunal" doesn't seem to find a problem using statistics or science to make their decisions.

http://www.manhattan-institute.org/html/cjm_35.htm

Serious scientific evidence is always enough to convict criminals.

[forgethell]

1. But 'serious scientific evidence' is

not the same thing as 'statistics'. Statistics is a method that can be used to arrive at scientific truth, but also to deceive. Remember the three kinds of lies: lies, damned lies, and statistics.

For instance, a non-random sample (for instance, a sample selected by someone with something to prove, although there could be many other reasons) would give a false conclusion.

Hard evidence is needed.

[libertypirate]

2. Hey does anyone know...

if there are any Vegas odds on whether * stole another?

[madison2000]

3. I worked as a technician in clinical research for years

I don't know how many meetings I have sat in where investigators attacked each others statistics- too small a sample, not randomized, you didn't perform such and such a test. Medical science depends upon statistics. We trust the whole medical establishment only to the degree that we trust statistical methods to be used properly to accurately represent the truth.

Suddenly, where votes are involved, its considered too soft to be evidence? We're talking about pollsters who know how to do this. We're talking about asking people about something they did less than an hour before who have no reason not to tell the truth. Do they think the results are made up? Biased?

And sometimes criminals are even convicted on totally circumstantial evidence, like Scott Peterson was recently.

Every single decision in the Bush V Gore debaucle was partisan, however. Its not about the law when it involves elections. Its about the outcome. Its about the political future of the judge.

[The Judged]

4. Actually, exit polls are sufficient: see Ukraine 2004

Also, it is in vogue to ask for a re-vote where the * candidate has lost an election, but not where the * candidate has won, given the same conditions but opposite political affiliations.