玉米区域试验中误差方差的异质性及其对品种评价的影响

Heterogeneity of Error Variance and Its Effect on Variety Evaluation in Corn Regional Trials

为了研究我国玉米区域试验中误差方差的异质性存在状况及其对品种评价的作用,以2003—2006年东北和华北16组玉米区域试验资料为依据,对玉米区域试验各环境试验误差方差差异状况及误差方差同质模型和异质模型的拟合效果进行了验证,并对品种效应差异显著性测验在误差同质模型和误差异质模型分析结果的差异状况进行了比较。结果表明,在分析的所有试验中,试验误差方差在环境间具有较大差异;误差方差异质模型比误差方差同质模型对试验数据拟合效果普遍较好;模型是否考虑误差方差的异质性对品种-环境交互效应测验结果有较大影响,而对品种主效应测验结果影响极小;误差方差异质模型比误差方差同质模型测验效率高。

The objective of this study was to improve the method for statistical analysis in corn regional trials by insecting error variance. Based on the 16 data sets of the corn regional trials in the Northeast and North China from 2003 to 2006, we assessed the error variation between the environments and the fit-goodness of models for homogeneous and heterogeneous errors, and compared the statistical tests for trial effects from the two models. The results showed that the error variance largely varied be- tween environments in all of the considered trials. The model for heterogeneous errors fitted the trial data better than the model for homogeneous errors. Whether the heterogeneity of error variance was considered in the models considerably impacted the test about variety-environment interaction effects and little did the test about variety effects. The model for heterogeneous errors had higher test efficiency than the model for homogeneous errors.