[英]Robust ANCOVA in R with package WRS2
The function ancboot
from the WRS
package performs a one-way ANCOVA with one covariate using 20% trimmed means and bootstrapping: WRS
软件包中的ancboot
函数使用20%的均值和自举来执行带有一个协变量的单向ANCOVA:
ancboot(formula, data, tr = 0.2, nboot = 599, sm = FALSE, fr1 = 1, fr2 = 1,
pr = TRUE, pts = NA)
The command returns an object of class ancova
containing trimmed means at 5 design points: 该命令返回5个设计点上包含修剪均值的ancova
类对象:
[1] "Note: confidence intervals are adjusted to control FWE"
[1] "But p-values are not adjusted to control FWE"
[1] "Taking bootstrap samples. Please wait."
$output
X n1 n2 DIF TEST ci.low ci.hi p.value
[1,] 2 15 30 -1.277778 -1.846494 -3.234577 0.6790212 0.0855
[2,] 4 38 50 -1.066667 -2.139515 -2.476449 0.3431158 0.0440
[3,] 4 38 50 -1.066667 -2.139515 -2.476449 0.3431158 0.0335
[4,] 5 42 50 -1.092308 -2.256915 -2.460882 0.2762670 0.0320
[5,] 6 25 41 -1.333333 -2.341798 -2.943341 0.2766743 0.0300
$crit
[1] 2.8277
However, I had expected to report the robust ANCOVA much like a conventional (parametric) ANCOVA, including an omnibus p value, 95% CI, F , and [df]. 但是,我曾希望像常规(参数)ANCOVA那样报告健壮的ANCOVA,包括综合p值,95%CI, F和[df]。 How can I get the desired output with a robust ANCOVA (ie, with trimmed means and bootstrapping)? 如何通过健壮的ANCOVA(即经过调整的方法和自举程序)获得所需的输出?
I checked the raw source file for any trace of methods, but it would appear the result has no special class, which means no special method can be defined. 我检查了原始源文件中是否有任何方法痕迹,但结果似乎没有特殊的类,这意味着无法定义特殊的方法。 Ask package author/maintainer to include this feature or contribute the source code. 要求软件包作者/维护者包括此功能或提供源代码。
声明:本站的技术帖子网页,遵循CC BY-SA 4.0协议,如果您需要转载,请注明本站网址或者原文地址。任何问题请咨询:yoyou2525@163.com.