01122017 11:16 AM
Hello, I am trying to run proc GLM considering disease severity on 102 samples and 3 environments. It is ok when I am considering only the variables by itself, but SAS does not give me pvalues when I try the interaction between them. Why is that? Is it because my model is saturated? How can I have that interaction analysed?
proc glm data=sev;
class line env;
model sev = line env line*env;
run;
Thank you.
01122017 02:32 PM
You only have 305 valid observations but ENV has 102 unique levels. Thus the model with the interaction term contains more parameters than there are observations.
Some options are
1. Get more data
2. Do not model interactions
3. Consider whether you can model ENV as a continuous variable
01122017 01:05 PM  edited 01122017 01:11 PM
Please show the output.
Also, please use PROC FREQ to provide the answers for
1. How many levels for LINE?
2. How many joint levels for ENV*LINE?
Since you didn't specify the SOLUTION option, I assume you are asking about pvalues in the Type 1 and Type III model ANOVA tables.
As the following example shows, PROC GLM does give pvalues for those tables when the data are nondegenerate:
proc glm data=sashelp.bweight plots=none;
class boy MomEdLevel;
model weight = boy MomEdLevel boy*MomEdLevel;
run;
01122017 02:22 PM
Hello,
Thank you for your answer.
Proc Freq gaves 3 levels for line and total of 306 env*line. I am looking for Type III model anova tables.
SAS Output
The SAS System 
102  1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 
3  1 2 3 
306 
305 
The SAS System 
304  67164.16118  220.93474  .  . 
0  0.00000  .  
304  67164.16118 
1.000000  .  .  34.64164 
101  30175.75618  298.76986  .  . 
2  5588.61953  2794.30976  .  . 
201  31399.78547  156.21784  .  . 
101  30089.07898  297.91167  .  . 
2  5588.61953  2794.30976  .  . 
201  31399.78547  156.21784  .  .

01122017 02:32 PM
You only have 305 valid observations but ENV has 102 unique levels. Thus the model with the interaction term contains more parameters than there are observations.
Some options are
1. Get more data
2. Do not model interactions
3. Consider whether you can model ENV as a continuous variable
01122017 04:13 PM
Thank you very much.
Need further help from the community? Please ask a new question.