[R] GLM Help

[peter dalgaard]

I think you are looking for

~ Region + Region:Helpers - 1

a.k.a. 

~ Region/Helpers - 1

Notice that these are actually the same model as your glm3 (and also as ~Region*Helpers), only the parametrization differs. The latter includes an overall Helpers term so that the interaction coefficients should be read as differences in slope. (With default treatment contrasts, the Helpers term would be the slope for the first region and the interactions are differences in slope compared to the first region).

-pd

On 03 Sep 2014, at 17:17 , Kathy Haapala <kathy at haapi.mn.org> wrote:

> Hi all,
> 
> I have a large set of data that looks something like this, although
> this data frame is much smaller and includes made up numbers to make
> my question easier.
> 
>> x.df <- data.frame(Region = c("A", "A", "A", "A", "A", "B", "B", "B", "B", "B", "B", "C", "C", "C", "C"), Group_ID = c(1:15), No_Offspring = c(3, 0, 4, 2, 1, 0, 3, 4, 3, 2, 2, 5, 4, 1, 3), M_Offspring = c(2, 0, 2, 1, 0, 0, 1, 1, 2, 0, 1, 3, 2, 1, 1), F_Offspring = c(1, 0, 2, 1, 1, 0, 2, 3, 1, 2, 1, 2, 2, 0, 2), No_Helpers = c(5, 0, 2, 1, 0, 1, 3, 4, 2, 3, 2, 3, 4, 0, 0))
> 
>> x.df
>   Region Group_ID No_Offspring M_Offspring F_Offspring No_Helpers
> 1       A        1            3           2           1          5
> 2       A        2            0           0           0          0
> 3       A        3            4           2           2          2
> 4       A        4            2           1           1          1
> 5       A        5            1           0           1          0
> 6       B        6            0           0           0          1
> 7       B        7            3           1           2          3
> 8       B        8            4           1           3          4
> 9       B        9            3           2           1          2
> 10      B       10            2           0           2          3
> 11      B       11            2           1           1          2
> 12      C       12            5           3           2          3
> 13      C       13            4           2           2          4
> 14      C       14            1           1           0          0
> 15      C       15            3           1           2          0
> 
> I have been using GLMs to determine if the number of helpers
> (No_Helpers) has an effect on the sex ratio of the offspring. Here's
> the GLM I have been using:
> 
>> prop.male <- x.df$M_Offspring/x.df$No_Offspring
>> glm = glm(prop.male~No_Helpers,binomial,data=x.df)
> 
> However, now I'd like to fit a model with region-specific regressions
> and see if this has more support than the model without
> region-specificity. So, I'd like one model that generates a regression
> for each region (A, B, & C).
> 
> I've tried treating No_Helpers and Region as covariates:
>> glm2 = glm(prop.male~No_Helpers+Region-1,binomial,data=x.df)
> which includes region-specificity in the intercepts, but not the
> entire regression,
> and as interaction terms:
>> glm3 = glm(prop.male~No_Helpers*Region-1,binomial,data=x.df)
> which also does not give me an intercept and slope for each region.
> 
> I'm not sure how else to adjust the formula, or if the adjustment
> should be somewhere else in the GLM call.
> 
> Thanks in advance for your help.
> 
> ______________________________________________
> 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.

-- 
Peter Dalgaard, Professor,
Center for Statistics, Copenhagen Business School
Solbjerg Plads 3, 2000 Frederiksberg, Denmark
Phone: (+45)38153501
Email: pd.mes at cbs.dk  Priv: PDalgd at gmail.com