R help - Sep 2011 - binomial logistic regression question

Dear subscribers, 

I am looking for a function which would allow me to model the dependent
variable as the number of successes in a series of Bernoulli trials. My data
looks like this

ID  TRIALS     SUCCESSESS   INDEP1 INDEP2 INDEP3
1      4444        0                0.273     0.055  0.156
2      98170      74               0.123     0.456  0.789
3      145486    30               0.124     0.235  0.007
4      147149    49               0.888     0.357  0.321
5      60585      11               0.484     0.235  0.235
6      198953    43               0.295     0.123  0.856

I want to find out how independent variables influence the number of
successful trials (dependent variable)
I had tried to use glm formula 

regression.glm <- glm( SUCCESSESS ~ INDEP1 + INDEP2 + INDEP3, data = data,
family = binomial, weights= TRIALS) 

but got the following: Error in eval(expr, envir, enclos) : y values must be
0 <= y <= 1
but there y has to be between 0 and 1 (which does make sense, but that's why
I am using weights?)

I would be grateful for any hints/suggestions on how to proceed.
Cheers, 
Juta


--
View this message in context:
http://r.789695.n4.nabble.com/binomial-logistic-regression-question-tp3846954p3846954.html
Sent from the R help mailing list archive at Nabble.com.

On 09/27/2011 07:53 AM, majesty wrote:
> Dear subscribers,
>
> I am looking for a function which would allow me to model the dependent
> variable as the number of successes in a series of Bernoulli trials. My
data
> looks like this
>
> ID  TRIALS     SUCCESSESS   INDEP1 INDEP2 INDEP3
> 1      4444        0                0.273     0.055  0.156
> 2      98170      74               0.123     0.456  0.789
> 3      145486    30               0.124     0.235  0.007
> 4      147149    49               0.888     0.357  0.321
> 5      60585      11               0.484     0.235  0.235
> 6      198953    43               0.295     0.123  0.856
>
> I want to find out how independent variables influence the number of
> successful trials (dependent variable)
> I had tried to use glm formula
>
> regression.glm<- glm( SUCCESSESS ~ INDEP1 + INDEP2 + INDEP3, data =
data,
> family = binomial, weights= TRIALS)
Try

regression.glm<- glm(cbind(SUCCESSESS,TRIALS) ~ INDEP1 + INDEP2 + 
INDEP3, data = data,family = binomial)

as is specified in the details of ?glm.

-- 
Patrick Breheny
Assistant Professor
Department of Biostatistics
Department of Statistics
University of Kentucky