Page 1 of 1

What is statistical significance? Supreme court takes it on.

Posted: Sat Apr 02, 2011 11:05 am
by MSI
April 2, 2011: Making a Stat Less Significant
  • A group of mathematicians has been trying for years to have a core statistical concept debunked. Now, the Supreme Cout might have done it for them.
    Last month, the court considered a case brought by investors in Matrixx Initiatives Inc. They alleged the company failed to disclose material information by neglecting to reveal it had received reports an over-the-counter medicine, Zicam Cold Remedy, caused a loss of sense of smell.
    When those reports came to light, the company's stock fell. Eventually, the Food and Drug Administration warned consumers not to use certain Zicam products
    The company argued in court that the initial reports of the possible side effect didn't rise to the level of "statistical significance," and therefore didn't need to be disclosed.
    A point on which most statisticians agree is that "statistical significance" is difficult to explain
see the full article: Making a Stat Less Significant
and be sure to see the blog A Statistical Test Gets Its Closeup

Re: What is statistical significance? Supreme court takes it on.

Posted: Tue Apr 05, 2011 4:47 pm
by brian
And on another related note about 'statistical' stuff at the Supreme court...
April 3, 2011: NY Times: When a Lawsuit Is Too Big
  • Can a class-action lawsuit be too sprawling to deliver old-fashioned justice?
    These types of cases routinely rely on statistical formulas rather than testimony and personnel records to decide how much money the company would have to pay each plaintiff if it lost.
    Judge Richard A. Posner of the United States Court of Appeals for the Seventh Circuit in Chicago was concerned about the due process rights of so-called absent class members: women who would be bound by the result in the case even though they had not volunteered for it and had no right to opt out and sue on their own. And he added that something more than a “mechanical computation” was needed to decide who gets how much money should the women win. Rather, he said, the case would require “500 separate hearings.”
see the full article: When a Lawsuit Is Too Big