I need SAS codes for statistical analysis of field experiments (split-plot design: A/(B*C)-R). Are there any differences between one year analysis and the analysis over several years (A/(B*C*D)-R)? It wasn't a static experiment. In the worst-case scenario, which non-parametric post-hoc test can I use? Is the "HOLM-procedure for contrasts" a suitable method? I already used it for block-designs.
Yes, there is difference between analysis of one year design and designs carried out over several years. I have the SAS codes that you want, but please be specific as much as you can, Explain about, Main Plots, Sub plots, was it Split Factorial? or Factorial Split?
We used a randomised split-plot-design (A/(B*C)-R) with 4 repetitions (R). Main plot was the date of harvest (A) und sub plots where variety (B) and N-fertilisation (C). Each block had another randomisation. Inside the blocks the main plots were randomised and each main plot had the same sub plot randomisation. The trial was carried out over three years.
in fact this is a Factorial design (subplots) repeated over time (Harvest time, main plots) which has been carried out across three years.
So you have five factors as follows:
Year (Random)=Y
Block (Random)=R
Harvest (Fixed)=H
Variety (Fixed)=V
Fertilizer (Fixed)=F
this is a very complicated design, the below SAS code should work
Proc GLM Data=(Your data);
Class Y R H V F;
Model Trait= Y R(Y) H R*H V F V*F H*V H*F H*V*F R*V R*F R*V*F Y*V Y*F Y*V*F Y*H*V Y*H*F Y*H*V*F Y(R*V) Y(R*F) Y(R*V*F)/ss3;
Random Y R(Y) R*H R*V R*F R*V*F Y*V Y*F Y*V*F Y*H*V Y*H*F Y*H*V*F Y(R*V) Y(R*F) Y(R*V*F);
Test H=R(Y) H e=R*H;
Run;
Quit;
please note that, Fisher Test for each of the factors should be applied according to the mathematical expectation of the mean squares.
Pleas test it and tell me about the results, if that was ok, i will explain you how to compare means.
Best Regards
I would probably use JMP from SAS. The programmers they have for JMP are top notch and make all of your data analysis for Split Plots really easy. They also have defaults for testing normality and homogeneity.
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics