[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Scientific method



If you want to compare two treatments you must supply replicates for EACH 
treatment. The question you then ask is whether the two means are different 
from eachother (this of course implies that you are measuring some 
parameter of importance). The two sets of data are then compared 
statistically 
to determine if the values within one treatment are significantly 
different from the values within the second treatment. You do this using 
what is called a t-test. You can also use what are called 
'non-parametric' statistics if you want to simplify your design.

You cannot test two treatments if there are no replicates because you 
have no 'degrees of freedom' which is another way (kind of) of saying 
that you dont know the variability within a treatment and therefor you 
dont know if it is 'different' than any other treatment.

dave.