I executed a study as two factorial experiments, as a randomized complete block design in three repetitions. The first experiment was executed using a N chemical fertilizer and phosphorus chemical fertilizer (as a 3×3 design with nine treatments) and the second experiment was conducted using a Nitroxin biofertilizer and biophosphorus biofertilizer (as a 3×3 design with nine treatments) ... The question is how can compare the treatments between these two factorial experiments, which test should I use in SAS or MSTAT-C?
I'm no statistician, so I would suggest that
- IF your experiments did not measure the same parameters (Yield? Leaf area? Nutrient uptake? - you are a little vague on the details for anyone to properly comment), and
- the experiments were not conducted at the same time (simultaneously or the following season?),
- under the same light, water and temperature conditions (pot trial? field trial? glasshouse?, again no details provided), and
- the Nitrogen and Phosphorus doses were not the same (again, no details just two 3x3 experiments).
THEN, you are unlikely to find a statistical test that will adequately account for all the confounding factors. I'm sure the software can run something to compare the results but there are so many caveats and assumptions that are likely to be broken, you might find it hard to justify to a reviewer.
Why not just graph the results? If it needs statistics to "prove" your results, perhaps there isn't enough in the data?
A few comments:
Three replicates is pushing the limits for adequate confidence limits in any plant growth trial. Did you consider the power of your experimental design? How much of an effect size will be required across this experimental design to be considered significant? It sounds like your experimental design is AxB vs CxD, where A and C are the nitrogen treatments and B and D the phosphorus. Perhaps you can reduce the experimental design to N x P treatments?
My apologies if this reply seems harsh. My experience with statisticians is that they are much friendlier to talk to before the experiment is performed, not afterwards.
Thank you for the answer
These 2 experiments were performed in the same block and type of soil and same time as well. both executed in 3 replication and 3 rates of fertilizer application for each fertilizer. chemical N rates ( 0, 45, 90 kg/ha) and Chemical phosphorous (0, 40 , 80 kg/ha).For biofertilizer treatments used Nitroxin in 3 rates (0 , 0.5, 1 L/ha) and Biophosphorous (0,0.5 ,1 L/ha). Can I use Contrast statement in SAS program to compare treatments between these two experiments?
- Badges
- Science topic