在农业科学试验中用图示法代替方差分析的尝试

利用W·A·Shewhart的质量控制图对农业科学试验结果进行分析,探讨用以代替方差分析的可行性.此法是在平面座标系中,根据试验结果的目标特性值,绘出三条具有统计意义的控制线.将样本平均值描在控制图上,如果目标值点子波动于控制界限之内,则表示该因素对目标特性值的影响不显著;如果点子超出控制界限之外,则表示有显著差异.用此法对二十余个试验结果进行了分析,其结果与方差分析结果完全一致.此法不仅易于被多数人掌握,而且简化了计算手续.

An Attempt at Using Diagram Instead of Analysis of Variance in Agricultural Experiments

According to the theory of mathematical statisties, the possibility of W. A.Shewhart's the diagram of the quality control used to analyse results of agricultural experiments instead of analysis of variance is studied.On the basis of the aim characteristic values of the experimental results,three control lines of statistics were designed, a central line, a upper control limit, a lower control limit in right angle coordinates. Then the average of samples was drawed on the control diagram. If the point of the average is inside control region, the factor would not influence the aim characteristic value significantly. If the point is outside control region it would have significant influence. This method was used over twenty experiments, the results were equle to that of analysis of variance. The method is not only easy to learn but also simply to calculate.