NavList:
A Community Devoted to the Preservation and Practice of Celestial Navigation and Other Methods of Traditional Wayfinding
Re: "Vernier acuity" of horizon IC tests
From: Nicol�s de Hilster
Date: 2009 Jul 10, 13:30 +0200
From: Nicol�s de Hilster
Date: 2009 Jul 10, 13:30 +0200
Bill Morriswrote: > Loosely speaking, isn't the standard error of the mean at a given level an indicator of the probability that the _mean_ lies within a specified range? We would expect the distribution of the means to be smaller than the distribution of the samples. > > I took n to be 30, rather than 30-1, as 30 seemed to qualify as a "large sample." The 1.96 gives us exactly 95%, versus your 2 SD and 95.45% (in case anyone is confused about where the figures come from). > > You seem to be writing about the probability that a given _observation_ lies within a specified range. > Hmmm, I should read more carefully next time (must have been my eye-sight), I was triggered by the 95% and thought you were talking about Standard Deviation, but you did mention Standard Error indeed. Sorry about that and for causing any confusion here.... As far as I remember the calculation - (1.96 x SD/root n) - should be done using the Sample Standard Deviation (which is the one where SD is calculated using the "n minus one" division), but that is just nitpicking in this case (the range will become slightly larger by a factor of app. 1.068, but despite that the outcomes will still not overlap). Nicol�s --~--~---------~--~----~------------~-------~--~----~ NavList message boards: www.fer3.com/arc Or post by email to: NavList@fer3.com To , email NavList-@fer3.com -~----------~----~----~----~------~----~------~--~---