My question is whether the distribution curve normalized to the total sample size? I mean for histograms, the y value usually is count in the bin. Different numbers of total sample will give different counts (or heights on y axis) even when the distribution pattern is the same. How can we automatically normalize the counts in each bin to total counts? Is the histograms/distribution curve in the 'Marginal Distribution Curves' already normalized to the total sample size?
The purpose of what I am trying to do is to compare the distribution of two group of data which have different sample size.
You are right, the height on Y axis of distribution curve will be different if you set different Bin Size.
If you want compare the distribution of two group of data, you need set same bin size for these two groups of data. There is no automatic way. You need change the binning setting one by one manually.