Extract standard errors from glm

Question

I did a glm and I just want to extract the standard errors of each coefficient. I saw on the internet the function se.coef() but it doesn't work, it returns "Error: could not find function "se.coef"".

Might help to put up some data and example code. – screechOwl Dec 13 '11 at 21:01 — screechOwl, Dec 13 '11 at 21:01

score 41 · Answer 1 · answered Dec 13 '11 at 21:11

The information you're after is stored in the coefficients object returned by summary(). You can extract it thusly: summary(glm.D93)$coefficients[, 2]

#Example from ?glm
counts <- c(18,17,15,20,10,20,25,13,12)
outcome <- gl(3,1,9)
treatment <- gl(3,3)
print(d.AD <- data.frame(treatment, outcome, counts))
glm.D93 <- glm(counts ~ outcome + treatment, family=poisson())

#coefficients has the data of interest
> summary(glm.D93)$coefficients
                 Estimate Std. Error       z value     Pr(>|z|)
(Intercept)  3.044522e+00  0.1708987  1.781478e+01 5.426767e-71
outcome2    -4.542553e-01  0.2021708 -2.246889e+00 2.464711e-02
outcome3    -2.929871e-01  0.1927423 -1.520097e+00 1.284865e-01
treatment2   1.337909e-15  0.2000000  6.689547e-15 1.000000e+00
treatment3   1.421085e-15  0.2000000  7.105427e-15 1.000000e+00

#So extract the second column
> summary(glm.D93)$coefficients[, 2]
(Intercept)    outcome2    outcome3  treatment2  treatment3 
  0.1708987   0.2021708   0.1927423   0.2000000   0.2000000

Take a look at names(summary(glm.D93)) for a quick review of everything that is returned. More details can be found by checking out summary.glm if you want to see the specific calculations that are going on, though that level of detail probably is not needed every time, unless you <3 statistics.

Are the standard errors stored within the `glm.D93` object? I couldn't eyeball it using `str()`. Or does `summary()` explicitly calculate the errors? — mindless.panda, Dec 14 '11 at 12:40
@mindless.panda - AFAIK they are calculated directly by `summary.glm`. If you type the function into your console sans `()` and then scroll down about 25 lines, you'll see where it's calculated. — Chase, Dec 14 '11 at 15:12

score 31 · Answer 2 · answered Dec 13 '11 at 21:24

31

Another way:

sqrt(diag(vcov(glm.D93)))

answered Dec 13 '11 at 21:24

Wojciech Sobala

7,431
2
21
27

This is beautiful - pure and simple. – zx8754 Mar 07 '14 at 10:18

score 6 · Answer 3 · answered Apr 17 '14 at 19:14

6

se.coef() actually does work. But it's not in the base package: it's in the {arm} package: http://www.inside-r.org/packages/cran/arm/docs/se.ranef

answered Apr 17 '14 at 19:14

Joel Chan

61
1
2

Extract standard errors from glm

3 Answers3

Linked

Related