CODEWITHSUNDEEP
rglm
Christina
Christina

Reputation: 9

GLM output trouble

So, i want to run a repeated measures GLM in RStudio, which I have done...for the most part.... However, Not all dates are showing up in my output (12/1/2015 is missing). Here is part of the output along with my model code so you can see what i mean:

CH4f1 <- glm(GC_CH4.flux~River*Site*Date*Hum.Hol, data = Rdata_w.o_OL_Date, family = gaussian)
 summary(CH4f1)

Call:
glm(formula = GC_CH4.flux ~ River * Site * Date * Hum.Hol, family = gaussian, 
    data = Rdata_w.o_OL_Date)

Deviance Residuals: 
    Min       1Q   Median       3Q      Max  
-37.307   -2.655   -0.341    0.314  163.247  

Coefficients:
                                 Estimate Std. Error t value Pr(>|t|)  
(Intercept)                       15.5643    75.6869   0.206   0.8372  
River                             -8.4098    55.1093  -0.153   0.8788  
Site                              -3.4776    25.9063  -0.134   0.8933  
Date12/1/2016                    112.5939    96.6623   1.165   0.2451  
Date4/1/2016                     -15.4780    96.9954  -0.160   0.8733  
Date4/1/2017                     -13.8752    94.5132  -0.147   0.8834  
Date6/1/2016                      12.5824    93.8721   0.134   0.8935  
Date6/1/2017                     -18.3304    94.5132  -0.194   0.8464  
Date9/1/2016                     170.2484    95.5697   1.781   0.0759 .
Date9/1/2017                     -38.4031    96.7184  -0.397   0.6916

How do I get 12/1/2015 to be one of the dates that shows up in my GLM output?

Upvotes: 0

Views: 230

Answers (1)

tmastny
tmastny

Reputation: 417

The column Date is a factor, or is being convert to a factor by glm.

By default, glm uses the treatment contrast for factors, also called Dummy Encoding. This means that (Intercept) is the coefficient for the first level, 12/1/2015.

For the other dates, those coefficients are changes from 12/1/2015. For example, the intercept for 6/1/2017 is actually (Intercept) + Date6/1/2017 = 15.5643 + ( -18.3304).

I could give more specific help if you provided a reproducible example.

Upvotes: 2

Related Questions