Hi and Good day,
I am in between a fix with my analysis. I joined a group with their experiment, where they have 3 blocks (A, B, C) with each block having two main treatment (X and Y), and each treatment having about 20 varieties of plants, we collected data for several days (DAP).
My question exactly is I am quite new to JMP, and finding it difficult to further my test, especially going ahead with an ANOVA.
What is the best approach for this sort of split plot design. Do I do an ANOVA per day ? Actually, when I do, it throws an error see image attached (per DAP), I am thinking because of the block-treatment-species interaction.
I am actually stuck on how best to do such an analysis, I would appreciate some guidance. I have attached a sample data.
Thank you.
Accepted Solutions
I am not sure that this experiment uses a split-plot design because I do not see any restriction on randomization. Blocking is a particular aspect of the design. I assume that Block contributes a random effect to the response. It also appears that all of the factors are crossed.
I would not separate the data by DAP and perform isolated ANOVA. I would not cross the blocking factor with any other factor because that says you think that the fixed effects in the process vary across blocks.
The following analysis is reasonable, based on my interpretation of your experiment.
I attached your data table with scripts to reproduce what I did to define the model and the results I obtained.
I am not sure that this experiment uses a split-plot design because I do not see any restriction on randomization. Blocking is a particular aspect of the design. I assume that Block contributes a random effect to the response. It also appears that all of the factors are crossed.
I would not separate the data by DAP and perform isolated ANOVA. I would not cross the blocking factor with any other factor because that says you think that the fixed effects in the process vary across blocks.
The following analysis is reasonable, based on my interpretation of your experiment.
I attached your data table with scripts to reproduce what I did to define the model and the results I obtained.