One way is to use the histcounts command to get the counts, and then use the bar command to do the plotting (because it will do the side-by-side):
rng 'default'
data1 = randn(20,1);
data2 = randn(30,1);
data3 = randn(40,1);
data4 = randn(50,1);
edges = -4:1:4;
h1 = histcounts(data1,edges);
h2 = histcounts(data2,edges);
h3 = histcounts(data3,edges);
h4 = histcounts(data4,edges);
figure
bar(edges(1:end-1),[h1; h2; h3; h4]')
I did not take great care to properly line up the "edges" variable with the counts. There are some nuances with respect to exactly how the binning is done. I suggest reading the documentation for histcounts.
Best Answer
As I suggested in this answer, I think the best you can do is to get the bin counts using the histcounts command, and then use the bar command to plot:
Here is an example.
Just as an FYI, I do think there is good reason to avoid the side-by-side histogram, which is that the bins do not line up where the actual data are, so it can be misleading. I speculate that this is why MathWorks eliminated this option.