Laserfiche WebLink
UII'A IIIMINI QI NAINII AI ^I SI II/IICIS <br />U. Mnnlc Pascoe, E><cculrvc Uucnor <br /> <br />1~11N II;U LAN U It I~;(;I.AI11 A'CIUN <br />423 Gentennlal Budding. 7313 Sherman Street <br />Denver, Coloratlo 80203 Tel. (303) 839-3567 I <br />David C. Shelton <br />Director <br />MEMORANDUM <br />T0: Sue Mowry <br />FROM: M. S. Savage 81av~as <br />RE: H-G Statistical Analysis <br />DATE: February 11, 1981 <br />Formula n = 2 S2 Z2 <br />min <br />(dX)2 <br />This formula is appropriate however: <br />1. 1.28 as a value for Z is not appropriate for a 90% level. They should ' <br />be using t = 1.645 (two tailed test at 90% confidence). However, this <br />is the correct value for an 80% confidence two-tailed test which they , <br />should be using. <br />2. 0.2 is not the degree of confidence, it is the level of accuracy <br />desired and sf)ould be 0.1 in this instance, since they wish to . <br />estimate within 10% of the actual mean. <br />3. The formula for this case should be: <br />n _ <br />min <br />2(5)2(1.28)2 <br />(0.1)2(X)2 <br />4. If H-G is utilizing this formula to determine minimum sample size <br />for reference areas and premining areas they must address where the <br />value for S (standard deviation) comes from. In order to accurately <br />apply this formula, the value for S must be a pooled variance (standard <br />deviation) of the nature: <br />S2 - (nR 1)5X1 + (nY-1)SY1 <br />nX+ny - 2 <br />where: <br />nX - number of samples from premining area <br /> <br />