我正在重新运行用于全面测试的代码,并且得到的输出少于以前。我正在运行一个带有2个术语和一个交互的模型。我的数据如下:
type pop chemoheterotrophy
1 larvae Bibel 0.929379395
2 larvae Bibel 0.945769649
3 larvae Bibel 0.991292099
4 larvae Bibel 0.947357992
5 larvae Bibel 0.926360018
6 larvae Bibel 0.960807347
有两种类型的流行音乐和5种流行音乐。我的模型如下:
k <- matrix(c(28.53,17.72,31.07,7.95, 7.65, 28.53,17.72,31.07,7.95, 7.65),1)
test = glht(glm(chemoheterotrophy ~ type + pop + type*pop, data=boxdata, family=quasibinomial(link="logit")),linfct=k)
当我以前用summary()
运行它时,我会得到intercept
,pop
,type
和pop*type
的p值。现在我得到:
Simultaneous Tests for General Linear Hypotheses
Fit: glm(formula = chemoheterotrophy ~ type + pop + type * pop, family = quasibinomial(link = "logit"),
data = boxdata)
Linear Hypotheses:
Estimate Std. Error z value Pr(>|z|)
1 == 0 -5.107 18.154 -0.281 0.778
(Adjusted p values reported -- single-step method)
知道发生了什么吗?代码和数据保持不变,但是现在我得到了不同的输出。