Laserfiche WebLink
Two statistical tests were selected for the demonstration of vegetation population <br />homogeneity. The fast test selected was Bartlett's Test (Zar, 1974). This test <br />evaluates the hypothesis that all sample variances come from populations with identical <br />variances, which then indicates that the sample variances estimate the same population <br />variance. This equality of sample/population variances is considered a demonstration <br />of homogeneity. The second test selected was a single factor Analysis of Variance <br />(ANON) (Zar, 1974). This test evaluates the hypothesis that all sample means (in this <br />instance, vegetation cover values) are equal. If all sample means are equal, then the <br />test hypothesis that the sample means come from the same population is not rejected, <br />again demonstrating homogeneity. The ANON test is dictated in instances where more <br />than two sample means area to be compared at the same time. Both Bartlett's test and <br />ANON are robust statistical tests, whose underlying assumptions of normal populations <br />are satisfied for these demonstrations. <br />BARTLETT'S TEST (HOMOGENEITY OF VARIANCES) <br />The sample data for this demonstration are the vegetation cover data from the Northern <br /># I Mine Arrz- West Side (N#1 West), Northern # 1 Mine- Topsoil Stockpile Area <br />(N#I T/S), Refuse Area- Topsoil Stockpile Area (R. A. T/S), Refuse Area (R.A.), and <br />Rienau # 2 Mine Area (R#2). These data are presented in full in the vegetation data <br />section of [his submittal. <br />Bartlett's Lest statistic is: <br />B= 2.3026((log SZp)(E vJ - (E v; log SZ;1 <br />Since the distribution of B appmzimates the X-square distribution, with v degrees of <br />freedom, a more accurate x-square estimate is obtained with the following correction <br />factor (Zar, 1974). <br /> <br />The correct test statistic then becomes: <br />Bc=B/C <br />