[R] error message in logistic regression

[David Winsemius]

On Mar 30, 2012, at 8:12 AM, carlb1 wrote:

> Hi
> I am trying to do a logistic regression on a small data file yet  
> when i get
> up to plotting the first set of graphs instead of 4 I only get one  
> graph,
> and some error messages. Yet it still looks like the program is doing
> something due to the "blue wheel" of the mouse.  Below is the script  
> copied
> from R and where I get up to before the message occurs.  The data  
> set is
> based on how accurate dogs  are at giving a trained response when  
> finding
> drugs.
> D=dogs, P= phase of test, NF= number found, NTA= times they gave the  
> trained
> alert, NA= times they did not give the alert, R= reliability in %
>
>> file.choose()
> [1] "E:\\Rhome exam\\question4\\table1Q4.txt"
>> tort<-read.table("E:\\Rhome exam\\question4\\table1Q4.txt")

At this point my brain said to itself... "this looks like a take-home  
exam problem. I wonder what the academic policy is at this person's  
institution about soliciting online help during exams?"

-- 
David.

>> tort
>   D  P NF NTA Na  R
> 1  1 1  27   8      19  30
> 2  1 2 29   7        22 24
> 3  2 1 30  16      14 53
> 4  2 2 26  25       1 96
>> summary(glm(cbind(NTA,NF-NTA)~P,tort,family=binomial))
>
> Call:
> glm(formula = cbind(NTA, NF - NTA) ~ P, family = binomial, data =  
> tort)
>
> Deviance Residuals:
>     1       2       3       4
> -1.339  -3.727   1.236   4.511
>
> Coefficients:
>            Estimate Std. Error z value Pr(>|z|)
> (Intercept)  -0.9671     0.6022  -1.606   0.1082
> P             0.6487     0.3830   1.694   0.0903 .
> ---
> Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
>
> (Dispersion parameter for binomial family taken to be 1)
>
>    Null deviance: 40.462  on 3  degrees of freedom
> Residual deviance: 37.556  on 2  degrees of freedom
> AIC: 54.505
>
> Number of Fisher Scoring iterations: 4
>
>> anova(glm(cbind(NTA,NF-NTA)~P,tort,family=binomial),test="Chisq")
> Analysis of Deviance Table
>
> Model: binomial, link: logit
>
> Response: cbind(NTA, NF - NTA)
>
> Terms added sequentially (first to last)
>
>
>     Df Deviance Resid. Df Resid. Dev Pr(>Chi)
> NULL                     3     40.462
> P     1   2.9064         2     37.556  0.08823 .
> ---
> Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
>> par(mfrow=c(2,2))
>> plot(glm(cbind(NTA,NF-NTA)~P*factor(D)+Na,tort,family=binomial))
> Error in qqnorm.default(rs, main = main, ylab = ylab23, ylim =  
> ylim, ...) :
>  y is empty or has only NAs
> In addition: Warning messages:
> 1: Not plotting observations with leverage one:
>  1, 2, 3, 4
> 2: In min(x) : no non-missing arguments to min; returning Inf
> 3: In max(x) : no non-missing arguments to max; returning -Inf
>>
>
>
> I have tried multiple ways of swapping and changing the factors  
> around in
> the equation yet I am still no further forward. Can anyone please  
> help and
> give me a hint as to where im going wrong?
>
> Much appreciated in advance
>
> --
> View this message in context: http://r.789695.n4.nabble.com/error-message-in-logistic-regression-tp4518865p4518865.html
> Sent from the R help mailing list archive at Nabble.com.
>
> ______________________________________________
> R-help at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.

David Winsemius, MD
West Hartford, CT