Laserfiche WebLink
STATE OP COLORArlO uu nrdui n. ~M. i~iw~.••~.,, <br />DEPARTMENT OF NATURAL RESOURCES <br />D. Monte Pascoe, Execu live D~, ector <br />1~11NEU LAND f{E(:LAI~IA'CION <br />423 Centennial Building, 1373 Sherman Street <br />Denver, Colorado 80203 Tel. i303) 639-3567 <br />David C. Shelton <br />Director <br />July 19, 1981 <br />Mr. David Duba <br />Morrison-Knudsen Company, Inc. <br />Two Morri son-Knudsen Plaza <br />P.O. Box 7808 <br />Boise, Idaho 83729 <br />Dear David: <br />Enclosed is a copy of my June 12, 1981 memorandum which addresses topics <br />which we discussed on June 10, 1981 with Brian Munson and Cass Legal. <br />I would reiterate that I have analyzed the H-G data in a manner irrespective <br />of the sample adequacy formula question. and found a large proportion of the <br />data to be below the minimum confidence levels N-G has committed to (August <br />5, 1980 letter from D. Duba to Carol Pahlke and Jim Herron). <br />Regarding the discussion on utilization of two-tailed versus one-tailed <br />analysis, I would submit that in the estimation of a parameter (for sample <br />adequacy purposes) utilizing the formula <br />t _ n (dx) 1 , <br />-~_ <br />it is appropriate to determine the level of confidence achieved by applying <br />the values from a one-tailed distribution. Since, in this instance, one is <br />constructing the confidence limits about the mean, a two-tailed application is <br />intrinsically implicated. Zn utilizing a one-tailed application, one <br />becomes concerned only with values which are above the upper critical value <br />(of the confidence interval), effectively accepting all values below this <br />critical point. This approach does not create a confidence interval about <br />the mean. <br />If you have any questions or comments, please contact myself or Brian Munson. <br />Sincerely, <br />Michael S. Savage <br />Reclamation Specialist <br />MSS/mmt <br />