number 51 here.
Vote in. Your project sounds well worth the exploration, and i wish you the best of luck in your scale-up phase.
I've been out of the statistics game for a while, but I remember enough to be happy for you--it's crucial that you have someone guiding you through the somewhat arduous task of manipulating the data to prove your hypothesis. When we design experiments our tunnel-vision can sometimes take over, and it's nice to have someone pushing you towards the "prove your conclusions" end of the spectrum; plus it allows for others to take your work WAY more seriously. There's nothing like numbers to settle an argument.
With that in mind, as I was looking at your latest plot (10 of each, treated vs non-treated), I found myself immediately thinking, is 10 enough to give you statistically significant results between the groups? I remember there is a simple test for that, just can't think of the name of it. I'm so happy that information has left my brain. Shoo, go away T-tests.
I hope the weather cooperates. Fingers crossed. No more overflowed creeks!
Cheers.
Gerardo
|