Laserfiche WebLink
1 <br /> I <br /> 1 <br /> samples for several organics methods. The same calculation <br /> { may be used for any test amenable to spiking: <br /> Where: <br /> P = Percent spike recovery <br /> A = Concentration determined on spiked sample <br /> B = Concentration determined on original uspiked sample <br /> T = True value of spike added <br /> Precision values may be calculated from analysis of dupli- <br /> cate nai.rs. In his manual Quality Assurance of Chemical <br /> Measurements, John K. Taylor of the National Bureau of <br /> Standards provides a formula for calculating the standard <br /> deviation based on a series of duplicates. When a <br /> sufficient number of spiked samples or duplicate pairs (at <br /> least 20) have been analyzed, control charts may be <br /> calculated. Because of the wide diversityof sample types <br /> and many different determinations, controp <br /> chartshave <br /> usually been calculated only in connection with .special <br /> ..' programs requiring them. However, the laboratory computer <br /> System is now being programmgd to calculate and maintain <br /> charts for every test on every .type of sample. For <br /> accuracy, the mean recovary and standard deviation (S) are <br /> calculated. Warning limits are set at +/- 2S and control <br /> limits at +/- 3S. For precision, since the least <br /> difference between a pair of duplicates is zero, that value <br /> forms the baseline of the chart.. The standard deviation <br /> provides upper limits, with the warning limit set at 2S and <br /> the control limit at 3S. <br /> t <br /> { <br /> VIII-4 <br /> J <br />