I am running logistic regression and want to be sure that I am getting 95% conficence intervals. Code:
# Dissable scientific notation.
# From: stackoverflow.com/questions/25946047
options(scipen=999)
###############################################################################
OR_CI_round_number<-5 # How many decimal point to keep after rounding OR and CI.
dfAPI <- haven::read_dta(
file = "https://stats.idre.ucla.edu/stat/stata/faq/eyestudy.dta")
dfAPI$carrot<-factor(dfAPI$carrot)
dfAPI$carrot<-relevel(dfAPI$carrot, ref = "1")
glmAPI = glm(lenses ~ carrot, data= dfAPI, family=(binomial(link = "log")))
#glmAPI
#summary(glmAPI)
round(exp(cbind(RR = coef(glmAPI), confint(glmAPI))), OR_CI_round_number)
round(exp(cbind(RR = coef(glmAPI), confint(glmAPI, level = 0.95))), OR_CI_round_number)
Result:
> round(exp(cbind(RR = coef(glmAPI), confint(glmAPI, level = 0.95))), OR_CI_round_number)
Waiting for profiling to be done...
RR 2.5 % 97.5 %
(Intercept) 0.41176 0.28349 0.54870
carrot0 1.58601 1.09250 2.40172
> round(exp(cbind(RR = coef(glmAPI), confint(glmAPI))), OR_CI_round_number)
Waiting for profiling to be done...
RR 2.5 % 97.5 %
(Intercept) 0.41176 0.28349 0.54870
carrot0 1.58601 1.09250 2.40172
The reason that I am asking is because I am getting RR 2.5 % 97.5 %. As I understand they indicate upper and lower boundaries of 95% Confidence intervals. Is this correct?
Yes, correct, those are your boundaries and areas to the left of 2.5% and to the right of 97.5% are significance levels.