Hello everyone,
I have a question regarding a statistic test.
I will give you the background first.
I am testing different characteristics of microglial cells. I am testing the difference between WT and KO mice. I have a sample of n=6. In each row, each genotype represents a mouse. Each value represents 1 cell.
If it is important - parameters from each mouse were taken from 4 fields (both hemispheres were imaged from 2 slices = 4 fields).
In the added table, I used nested t-test, as a representative from Prism suggested.
Do you think nested t-test is the best test for my DATA?
Is it legitimate to use this test in this experiment?
Ela.
As I see it from you description (same info attached as well):
>>>
I have a question regarding a statistic test.
I will give you the background first.
I am testing different characteristics of microglial cells. I am testing the difference between WT and KO mice. I have a sample of n=6. In each row, each genotype represents a mouse. Each value represents 1 cell.
If it is important - parameters from each mouse were taken from 4 fields (both hemispheres were imaged from 2 slices = 4 fields).
In the added table, I used nested t-test, as a representative from Prism suggested.
Do you think nested t-test is the best test for my DATA?
You are only comparing two conditions (WT vs KO) so a t-test (or non-parametric equivalent) is the proper choice of statistical test. If you are comparing the characteristics between individual cells of the two genotypes, then you can run the t-test on all of the cells from each genotype directly. If you are looking at something characteristic to the fields (e.g. cell counts/density) then you need to average the fields for each individual mouse and use those values for the comparison.
Either way, I'd recommend visualizing the data points from each mouse (i.e. make a column for each mouse) to make sure that the distributions are consistent from animal to animal within each genotype.