[R] Plot for Binomial GLM

[Joshua Wiley]

On Mon, Oct 4, 2010 at 7:21 AM, klsk89 <karenklsk89 at yahoo.com> wrote:
>
> Hi i would like to use some graphs or tables to explore the data and make
> some sensible guesses of what  to expect to see in a glm model to assess if
> toxin concentration and sex have a relationship with the kill rate of rats.
> But i cant seem to work it out as i have two predictor
> variables~help?Thanks.:)

What about xtabs?   For instance:

xtabs(deadalive ~ Dose + Sex, data = rat.toxic)

Regarding graphs, take a look at faceting in ggplot2 (or lattice).
You can get something close to the 3 way table but in graphical form
that way.  I am not sure if this is completely up and running yet, but
I know there has been work linking ggobi with R.  I have seen a few
demonstrations that looked quite promising, and it may work well for
you to visualize three variables at once (and interactively).  Here is
the link:  http://www.ggobi.org/rggobi/

>
> Here's my data.
>
>> rat.toxic<-read.table(file="Rats.csv",header=T,row.names=NULL,sep=",")
>> attach(rat.toxic)
      ^  why attach it?
>> names(rat.toxic)
> [1] "Dose"  "Sex"   "Dead"  "Alive"
>> rat.toxic
>   Dose Sex Dead Alive
> 1    10   F    1    19
> 2    10   M    0    20
> 3    20   F    4    16
> 4    20   M    4    16
> 5    30   F    9    11
> 6    30   M    8    12
> 7    40   F   13     7
> 8    40   M   13     7
> 9    50   F   18     2
> 10   50   M   17     3
> 11   60   F   20     0
> 12   60   M   16     4
> 13   10   F    3    17
> 14   10   M    1    19
> 15   20   F    2    18
> 16   20   M    2    18
> 17   30   F   10    10
> 18   30   M    8    12
> 19   40   F   14     6
> 20   40   M   12     8
> 21   50   F   16     4
> 22   50   M   13     7
> 23   60   F   18     2
> 24   60   M   16     4

Please tell me that after this, you converted the counts of dead and
alive into a single variable that had a 0 or 1 if dead and the
opposite as alive before you used it as the dependent variable in your
logistic regression.

> glm2<-glm(deadalive~Dose*Sex,family=binomial,data=rat.toxic)
>> anova(glm2,test="Chi")
> Analysis of Deviance Table
>
> Model: binomial, link: logit
>
> Response: deadalive
>
> Terms added sequentially (first to last)
>
>
>         Df Deviance Resid. Df Resid. Dev P(>|Chi|)
> NULL                        23    225.455
> Dose      1  202.366        22     23.090    <2e-16 ***
> Sex       1    4.328        21     18.762    0.0375 *
> Dose:Sex  1    1.149        20     17.613    0.2838
> ---
> Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
>> summary(glm2)
>
> Call:
> glm(formula = deadalive ~ Dose * Sex, family = binomial, data = rat.toxic)
>
> Deviance Residuals:
>     Min        1Q    Median        3Q       Max
> -1.82241  -0.85632   0.06675   0.61981   1.47874
>
> Coefficients:
>            Estimate Std. Error z value Pr(>|z|)
> (Intercept) -3.47939    0.46167  -7.537 4.83e-14 ***
> Dose         0.10597    0.01286   8.243  < 2e-16 ***
> SexM         0.15501    0.63974   0.242    0.809
> Dose:SexM   -0.01821    0.01707  -1.067    0.286
> ---
> Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
>
> (Dispersion parameter for binomial family taken to be 1)
>
>    Null deviance: 225.455  on 23  degrees of freedom
> Residual deviance:  17.613  on 20  degrees of freedom
> AIC: 91.115
>
> Number of Fisher Scoring iterations: 4
>
>
>
>
>
>
> --
> View this message in context: http://r.789695.n4.nabble.com/Plot-for-Binomial-GLM-tp2954406p2954406.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.
>



-- 
Joshua Wiley
Ph.D. Student, Health Psychology
University of California, Los Angeles
http://www.joshuawiley.com/