
Two means in one sample
Hi (Hi),
I'm not sure if this belongs in the basic bit but the question is pretty simple...
I have some biological orientation data in a histogram that looks like it has two normally dsitributed means in it, they are spread out with one mean a good distance from the other.
Does anyone know how to test for each of the peaks in the data?
Is there a mathematical test for this,
a) Number of peaks in the histogram. (I could do this in code)
b) It's normally distributed.(Based soley on 1 image)
I could just separate them in code and test each peak within it's range, but I'm not sure if that is mathematically sound? Or whether that's a statistical no no :eek:...
Apparently this is called a bipolar distribution, and I've been told it's ok to separate them, I'm still not 100% convinced though as it seems arbitrary.
Any help would be appreicated (Happy)

Clearly, there can't be two means.
I'm guessing that you have two modes.
http://en.wikipedia.org/wiki/Bimodal_distribution
It's called bimodal and not bipolar.
I know too many people that can best be described as bipolar.

Thanks for the reply, I know what to search for now...
Thanks for clarifying my mistakes, mode not mean, and bimodal not polar.
hehe it's no wonder I couldn't find it looking for bipolar (Giggle)
If you happen to know a good book for this area I'd appreciate it (Bow)