Message-ID: <50B8A196.6020905@sapo.pt>
Date: 2012-11-30T12:07:50Z
From: Rui Barradas
Subject: Confidence intervals for estimates of all independent variables in WLS regression
In-Reply-To: <CACm_P7rRJjVEZYj-imEo64N2KAQ_wQPEtOwLGa+0NL8r=PhqhQ@mail.gmail.com>
Hello,
You don't have to exchange 'object' by the name of your model, you call
the function with the name of your model:
x <- 1:20
y <- x + rnorm(20)
fit <- lm(y ~ x)
ci_lm(fit)
beta lower upper
(Intercept) 0.6741130 -0.9834827 2.331709
x 0.9575906 0.8192171 1.095964
Hope this helps,
Rui Barradas
Em 30-11-2012 01:07, Torvon escreveu:
> Rui,
>
> Thank you very much. Are there other things I have to adjust except for
> exchanging "object" by the name of my model?
> Torvon
>
> On 29 November 2012 08:17, Rui Barradas <ruipbarradas at sapo.pt> wrote:
>
>> ci_lm <- function(object, level = 0.95){
>> summfit <- summary(object)
>> beta <- summfit$coefficients[, 1]
>> se <- summfit$coefficients[, 2]
>> df <- summfit$df[1]
>> alpha <- 1 - level
>> lower <- beta + qt(alpha/2, df = df)*se
>> upper <- beta + qt(1 - alpha/2, df = df)*se
>> data.frame(beta, lower, upper)
>> }
>>