Laserfiche WebLink
3.0 STATISTICAL SUMh1ARY OF RESULTS <br />• Table 9 is a summary of mean, standard deviation, sample number, t-computed <br />and t-tabular for each of the data sets requiring additional sampling. <br />Section 2.0 describes the formula for computing t and for obtaining it from <br />the t-table., The summary table indicates that in all cases the computed t <br />value exceeds the tabular value, so that all data sets estimate within at <br />least 10% of the actual mean value with SO% statistical confidence. <br />One-tailed t-tests at the 95% confidence level were calculated for cover <br />and productivity with means for each vegetation type on reference and mine <br />areas. For woody plant density, t-tests were not performed since the <br />reference area data will not be used for determining revegetation success. <br />The null hypothesis was that the parameters measured were not significantly <br />different on the mine and reference areas. If the null hypothesis was <br />rejected, this would indicate the means were significantly different at the <br />chosen level of testing. If the measured reference area parameter was <br />greater than that occurring on the mine area, this was acceptable. Thus, <br />the null hypothesis could be rejected (means significantly different) with <br />• a satisfactory result if the reference area parameter was greater than the <br />mine area parameter. <br />The following formulas were used: <br />t- IX1-X2I <br />Sp nI + n2 and Sp (nI - I) SI2 + (nZ - 1)S22 <br />nlxn2 nl+n2-2 <br />Where R = mean values of two means being compared <br />Sp pooled variance (shown at right) for samples with equal <br />variances <br />S variance of each sample <br />t the computed t value <br />n = the number of observations in each sample <br />The computed t value was compared with the tabular t value aL.the 95% <br />• confidence level for a one-tailed test with the appropriate degrees of <br />63705 <br />