Multivariate analysis for environmental factors

Question

Hello there, a question by a novice.

I have data on the quantity of certain molecules in the essential oil of a plant I am researching. I've tested two factors, Temperature and Length of day to see how they affect the quantity of these molecules. There are 4 different temperatures, 2 different day lengths (a total of 8 treatments) and 18 molecules tested in 3 repeats of each treatment.

What would be the best test to apply to these results to see which treatment is correalated with certain abundance of a molecule, and how should I do that.

Thanks.

Extract.jmp

Mark_Bailey · Answer

Additionally, you could build a model for each response to explore the relationship between them and the conditions defined by the two factors. Run the table script in data table that I attached to see the individual fits and the combined response profiles. There are missing values that compromise some of the models but I think that they are all useful.

Several responses exhibit similar profiles and you might expect that knowing the biochemistry behind the extracts.

I must say that you are lucky. You set up a study, collected the data, and THEN ask how to analyze/model it? You mean, like "Fire, Ready, Aim!"? What guarantee did you have that the data would support the analysis that would provide the answers you sought? I do hope that your study is successful. I just want you to know that Design of Experiments begins with the model.

Extract.jmp

mlo1 · Answer

e.g. exploratory start using Graph builder

P_Bartell · Answer

If all you seek is examining for correlation (as opposed to modeling?), a graphical approach is always a good place to start. I'd just start with JMP's Graph Builder and start dragging and dropping factors and responses onto the graphing zone. Maybe mix in a local data filter for easy peeling back of the correlation onion. For a more analytic as well as graphical approach, the Multivariate platform is your place to start.