On Sun, 15 Feb 2004 13:58:32 GMT, "mejustme" <[EMAIL PROTECTED]> wrote:
> When you use statistics for experimental physics, you often find several > values with different absolute errors (supposed to be 100% interval). > I don't think that physics pretends to give 100% intervals. Do you have an example from a formal paper? I assume that in physics, as in other sciences, this is something that is likely to be in the footnotes. (I could imagine having to turn back to the original citation, but that seems like sloppy writing for an academic paper.) > What is the best way to obtain a value that does incalculate the errors? > (which are not standard deviations or variances or something likely) > > i thought about dividing them by 3, which is +- sigma (as 3*sigma gives a > 99% interval), and then using the mean value incalculating their weight) > -- Rich Ulrich, [EMAIL PROTECTED] http://www.pitt.edu/~wpilib/index.html "Taxes are the price we pay for civilization." . . ================================================================= Instructions for joining and leaving this list, remarks about the problem of INAPPROPRIATE MESSAGES, and archives are available at: . http://jse.stat.ncsu.edu/ . =================================================================
