There is fourteen locations of Guava orchard in the Jhalawar district. Each location is considered as treatment. Bearing tree will be purposely selected for this study.
Bhim,
It's not clear from your statement what kind of a study you are conducting. Are you surveying the differences among these orchards, as "treatment"? Or are you contemplating treating some locations with pruning/cultivating/covercropping/chemicals (etc.) and studying the effect (before and after, cost effectiveness, environmental impact, etc.)? If the latter, are you asking if you should select trees which are physico-chemically similar (i.e., re/ total soluble solids, total titrable acidity, fixed acidity, volatile acidity, pH, vitamin C, antioxidant profile, odor components, etc.) or as varied as possible, or neither (just representative of the orchard average)? If there is one treatment, to be applied or not, then representative trees (before and after) can serve as their own controls (successive years, different branches, etc.), with orchard location a second factor.
Bhim -
I am not an expert on your subject matter, but was wondering why the "bearing tree" is purposefully sampled? My work involved purposeful sampling, but it was because there was good modeling (regression prediction) data available, for highly skewed establishment surveys.
Often a model may greatly improve the "representativeness" of the results of a sample, especially in conjunction with a randomized sample design, and a randomized sampling method alone can be very helpful, but if you have neither randomization nor a model (both would be really good), that is problematic.
Perhaps you are looking for representativeness in the survey selection process and figure that you can do better for that kind of representativeness than selection at random. Perhaps you can, but that is not scientifically verifiable, and can lead to a biased selection and a biased result as well. There would be no measure of variance or bias to use, just perhaps a follow-up study. Your methodology needs to be quantifiable if you are to have useful results for such a survey.
As I said, I don't know your subject matter, but it certainly appears worthwhile. Perhaps you have subject matter related reasons for your tree selection? Perhaps you are only looking at a class of trees within each group, and you could randomly select among the eligible cases?
Best wishes on what sounds like an interesting project.
Cheers - Jim
- Badges
- Science topic
- Similar topics
- Agricultural Science
- Horticulture