Re: comparison of two distribution curves

From: john (jkalexan_at_gmail.com)
Date: 01/31/05 

Date: 31 Jan 2005 09:52:08 -0800

Thanks to everyone who has helped me get through this!

Here is what I have settled on ....

I have run a 2-factor ANOVA - which can tell me if the "conditions" are
different,
then I use a Bonferroni correct post-hoc to test each data point for
difference from all others,

more importantly, though, to compare between my experimentals,
I summarized each sample into a single number - the volume between 80
and 100 degrees.
This is assuming the peak is at 90 - which there is strong evidence
since the 2-factor ANOVA pionts out those are the significant points.
I can then compare the "volumes" between each condition. this looks
like it is going to work well.

Also, I think it is a good idea to smooth the raw data - i tried a 5
point moving average - which really made it look better - there are
plenty of reasons to do this - one reason why i avoided comparing
"peak" values is because, on occasion, an outlier at a single datapoint
would be higher than the obvious peak of the curve - which would
provide a "false" peak. the moving average solves this issue.
any problems i should be aware of with the moving average?

John

Relevant Pages