03-23-2017 05:18 AM
If it would be possible for you, please help me to find a solution for this problem. First I will explain the project completely. We are a greenhouse team and we try to find out the effect of irrigation and density and variety in green house. Irrigation need big plot and density need medium plot and variety need small plot. So we have to use split-split plot design in CRBD. Our problem starts from the effect of year and location in SAS. This project has done in woody and metal structure greenhouse in two years. In fact, woody and iron structure are the same in both years but we couldn’t write a suitable code for split-split plot design in 2years and 2locations. Please help us to solve this problem. With best regards Bardia
irrigation is main plot × Density is subplot × variety is sub subplot × 2 years × 2 different locations
03-23-2017 05:58 AM
Googling for "SAS CRBD" found me this (among other hits): https://www.ndsu.edu/faculty/horsley/SAS_for_sp_and_cmb.pdf
Could this be useful for you?
I also moved your post to the Statistic Procedures community.
03-23-2017 10:49 AM
First I have to say thank you for your attention. This file which you attach previously were completely use full. If it is possible for you, please help us to calculate hole of the project because in the attached file "SAS Commands for the Analysis of an RCBD with a Split-split Plot Arrangement" we can easily calculate our data but if we want to calculate the effect of location and year we have to use this code which didn’t contain three level (a b c) (irrigation is main plot (a) × Density is subplot (b) × variety is sub subplot (c))
SAS Commands for the Combined ANOVA Across Locations for an RCBD with Factorial Arrangement of A and B (locations are random and A and B are both fixed effects)
ods graphics off;
ods rtf file='cmb_loc_anova.rtf';
class rep a b;
model yield=rep a b a*b;
title 'ANOVA for Each Individual Location';
class loc rep a b;
model yield=loc rep(loc) a loc*a b loc*b a*b loc*a*b;
test h=a e=loc*a;
test h=b e=loc*b;
test h=a*b e=loc*a*b;
means a/lsd e=loc*a;
means b/lsd e=loc*b;
title 'Combined ANOVA Across locations Assuming Location is a Random Effect and A and B
are Both Fixed Effects';
in this code we only have (a) and (b) treatment but in fact we have (a) (b) (c) in our project and on the other hand this code were specialized for RCBD but our project is split-split RCBD . my first question is how could we calculate the effect of years and location on split-split RCBD. If we use the code of page 7 of attached file which show how to calculate split-split RCBD without the year and location, we have to calculate separately each location and each year? My second question is if we have to use the code in page 23, how we could calculate (c) because this code work with (a) and (b) and on the other hand code of page 23 is for RCBD but our project is split-split RCBD. If it would be possible for please help us. With best regard Bardia
03-23-2017 11:03 AM
I have to admit I'm a total dummy when it comes to statistics, I'm just a data engineer. But I know how to find information, so I attempted to point you where you could find more.
I would not be surprised if @Rick_SAS could lend a helping hand here.