NavList:
A Community Devoted to the Preservation and Practice of Celestial Navigation and Other Methods of Traditional Wayfinding
Re: Standard Deviation Question
From: Bill B
Date: 2013 Jan 04, 22:54 -0500
From: Bill B
Date: 2013 Jan 04, 22:54 -0500
On 1/4/2013 8:10 PM, William Hawes wrote: > Not being a math wiz, when should I be using stdDevPop and when should I > be using stdDevSamp? The TI-Nspire Reference Manual does not shed any > light on when to use the two different Standard Deviation functions. Can > anyone elaborate? This should be fun if the heavyweights join in. My math is limited to Purdue U. calculus and statistics as they apply to undergrad consumer/market research. Alex (a world-class mathematician) has posted results for relatively small samples in both n-1 (sample) and n (population). If I recall correctly I queried him personally as to what was a sample and what was a population and he suggested that it is a tough call. An example from a text (Statistics for Dummies) uses an example of taking a measurement of every child in a teacher's first-grade home room. Following what I believe Alex was trying to impart to me: It could be a population as it is all the kids in the home room (unless one was home sick). More than one first-grade home room in the school, maybe it is a sample of the school. Is it a sample of all the first-grade homerooms or all first graders in the nation? At least that's what I came away with--a slippery slope. Hopefully Alex will jump in and clear up my fuzzy interpretation. As per coin flips or sextant observations etc. there could be in theory be an infinite number so anything is just a sample. Personally I tend to treat 5 or 10 observation as a sample but run the n (population) anyway. The deviation is smaller and makes me feel all warm and fuzzy ;-) Bill B