Heatplot for Spearman correlation

[Maria Santos]

Hello all,
I am running some correlations and plotting a heatmap matrix, but I would like to know how can I add a star for significance level (<5%). I did not find something helpful. The command I am using is the following:

spearman var1 var2 var3 var4, stats(rho p) star(.05), if var5==1
matrix C = r(Rho)
heatplot C, legend(on) color(hcl diverging, intensity(.6)) keylabels(, format(%4.2f)) lower

Is there any addicional command to add a star to the significant correlations in my plot? Can someone help me in this regard?
Thanks in advance!

[Andrew Musau]

heatplot is from SSC (FAQ Advice #12). See https://www.statalist.org/forums/for...about-heatplot

[Maria Santos]

The solution I found was to add asterisc (*) in the edit graph option.

But generating the plots I run into another problem, it seems the gradient colours are not matching the significant values. For example, the white color still gives me a significative value for the Spearman correlation as seen in this picture:

I am not sure if I was clear. Please, help me!

[Andrew Musau]

You are starring based on the coefficients and not the p-values, which is not what you intend. I already referred you to the link in #2 and I see no evidence that you followed the procedure outlined there.

[Maria Santos]

You are starring based on the coefficients and not the p-values, which is not what you intend. I already referred you to the link in #2 and I see no evidence that you followed the procedure outlined there.

The star was added based on the p values generated from the spearman test. But what I don't understand is how my colours don't match the correlations. The white was supposed to be the coefficient closest to zero, but that is not working here. I would like to fix this cause I need to generate this graphs as soon as possible for my qualification exam :/

[Andrew Musau]

As recommended in FAQ Advice #12, you need to create a reproducible example to increase your chances of getting helpful replies. Otherwise, I have offered all the help that I can with no data example.