從 coeftest 獲得置信區間？

Question

有沒有辦法讓 coeftest 報告置信區間？ 還是一個從 coef 結果計算置信區間的命令？

虛擬數據：

library(lmtest)
data("mtcars")
testmodel <- glm(am ~ vs, data = mtcars, family = quasibinomial(link = "logit"))
testcoef <- coeftest(testmodel)
testcoef
z test of coefficients:

            Estimate Std. Error z value Pr(>|z|)
(Intercept) -0.69315    0.51640 -1.3423   0.1795
vs           0.69315    0.75593  0.9169   0.3592

Answer 1

那會有幫助嗎？

library(lmtest)
data("mtcars")
testmodel <- glm(am ~ vs, data = mtcars, family = quasibinomial(link = "logit"))
testcoef <- coeftest(testmodel)

confint(testcoef)
coefci(testmodel)

Answer 2

在@Will 之后，您可以使用cbind將coeftest與coefci結合起來

cbind(coeftest(testmodel), coefci(testmodel))
cbind(coeftest(testmodel), confint(testmodel))  ## profile CIs (more accurate)

或使用broom::tidy ：

library(broom)
tidy(testmodel, conf.int = TRUE)  ## profile CIs (the only option)

如果你有多個模型，使用purrr::map_dfr合並結果很方便：

data(mtcars)
m1 <- glm(am ~ vs, data = mtcars, family = quasibinomial(link = "logit"))
m2 <- update(m1, . ~ mpg)
(list(vs = m1, mpg = m2)
  |> purrr::map_dfr(tidy, conf.int = TRUE, .id = "model")
)