03-23-2018 02:37 PM
I am conducting a two-factorial anova analysis with only one observation per cell.
How can I run the proc glm statement? I will add an interaction term for the two factors: factor a* factor b;
Proc glm data=mydata;
class factor1 factor2;
model y= factor1 factor2 factor1*factor2; run;
I am currently getting having an issue and the error SS and MS is coming up as 0.00;
I need to also break down the error SS into residual and interaction parts.
03-23-2018 03:25 PM
This sounds like a data problem. What do you mean by "one observation per cell"?
How many observations in the data set? How many observations in each factor?
You can count the observations in each factor and in all by using the following call to PROC FREQ:
proc freq data=mydata; tables factor1*factor2; run;
03-23-2018 04:45 PM
Its not a data issue. This is a two factor design, and there are one observations within each factor combination.
How can I get ANOVA to display the SS and MS for each effect? It only give me the full model and error effects. I tried using the test statement as follows:
Test h=[effect of interest] e=??? ;
I am not sure what to put in the error statement, I just want it with the default error of MSE(mean squared error).
03-23-2018 08:26 PM - edited 03-23-2018 08:27 PM
Show us your output from PROC GLM. Show us the result of running PROC FREQ as suggested above by Rick.
Do not reply with words but not showing us your output. Show us both outputs. Not one of the two outputs requested, show us both.